Pith. sign in

REVIEW 1 major objections 2 minor 60 cited by

Reviewed by Pith at T0; open to challenge.

T0 means a machine referee read the full paper against a public rubric. The mark states how deep the mechanical check went, never who wrote it. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.3

A dataset of everyday scenes with per-instance segmentations advances object recognition in context.

2026-05-11 23:56 UTC pith:723G7MIG

load-bearing objection COCO is the dataset paper that delivered the scale and per-instance segmentations needed for realistic scene understanding benchmarks. the 1 major comments →

arxiv 1405.0312 v3 pith:723G7MIG submitted 2014-05-01 cs.CV

Microsoft COCO: Common Objects in Context

Tsung-Yi Lin , Michael Maire , Serge Belongie , Lubomir Bourdev , Ross Girshick , James Hays , Pietro Perona , Deva Ramanan
show 2 more authors
C. Lawrence Zitnick Piotr Doll\'ar
This is my paper
classification cs.CV
keywords object recognitionscene understandinginstance segmentationcrowd sourcingcommon objectseveryday scenesCOCO datasetDeformable Parts Model
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper presents the COCO dataset to advance object recognition by situating it within scene understanding. Images of complex everyday scenes are gathered and objects labeled with per-instance segmentations for precise localization. The collection includes 91 common object types in 328,000 images totaling 2.5 million instances, created via novel crowd-sourcing interfaces. Statistical comparisons to existing datasets and baseline detection results using Deformable Parts Models are included. A reader would care if this enables models to better handle real-world variability and context.

Core claim

The authors present a new dataset with the goal of advancing the state-of-the-art in object recognition by placing the question of object recognition in the context of the broader question of scene understanding. This is achieved by gathering images of complex everyday scenes containing common objects in their natural context. Objects are labeled using per-instance segmentations to aid in precise object localization. The dataset contains photos of 91 objects types that would be easily recognizable by a 4 year old. With a total of 2.5 million labeled instances in 328k images, the creation drew upon extensive crowd worker involvement via novel user interfaces for category detection, instanceスポ

What carries the argument

The per-instance segmentation labels on images of common objects in natural scenes, enabled by crowd-sourcing interfaces for detection, spotting, and segmentation.

Load-bearing premise

The novel crowd-sourcing interfaces produce sufficiently accurate and consistent per-instance labels at the reported scale without substantial systematic errors.

What would settle it

An independent audit revealing high disagreement rates between the provided labels and expert annotations on a sample of images would falsify the dataset's utility for advancing reliable recognition.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Object recognition models can be evaluated and trained on contextual scenes rather than cropped objects.
  • Per-instance masks support tasks requiring precise boundaries like segmentation.
  • Large scale supports statistical robustness in algorithm comparisons.
  • New annotation methods scale labeling to millions of instances efficiently.
  • Shifts focus toward holistic scene understanding in computer vision research.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This could improve transfer to applications like video analysis or robotics where context matters.
  • The dataset design may inspire similar collections for other modalities or languages.
  • One might test if models trained on COCO generalize better to unseen scenes than on prior datasets.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

1 major / 2 minor

Summary. The manuscript presents the Microsoft COCO dataset: 328k images of complex everyday scenes with 2.5 million per-instance segmentations across 91 object categories recognizable by a 4-year-old. The central goal is to advance object recognition by embedding it in scene understanding, achieved via novel crowd-sourcing interfaces for category detection, instance spotting, and instance segmentation. The paper supplies statistical comparisons against PASCAL, ImageNet, and SUN plus baseline Deformable Parts Model results for bounding-box and segmentation detection.

Significance. If label quality holds, the dataset would be a major resource for the field: its scale, instance-level annotations in natural contexts, and explicit comparisons to prior collections provide a reproducible benchmark that can drive progress on precise localization and contextual recognition. The manuscript's strength lies in the independent construction and public release of the data together with parameter-free statistical characterizations.

major comments (1)
  1. [Dataset construction / annotation interfaces] Section on dataset construction and crowd-sourcing interfaces (description of instance spotting and segmentation UIs): no inter-annotator agreement scores, no expert-vs-crowd IoU on a held-out subset, and no error analysis for boundary precision or category confusion are reported for the 2.5M instances. Because the central claim rests on these labels enabling precise localization in scene context, the absence of quantitative validation is load-bearing; systematic biases (e.g., under-segmentation of small or occluded objects) would directly undermine downstream utility.
minor comments (2)
  1. [Baseline performance analysis] The baseline DPM results are presented mainly as reference points; a short discussion of why more contemporary detectors were not included would clarify the intended role of these numbers.
  2. [Statistical analysis] Tables and figures comparing statistics to PASCAL/ImageNet/SUN would benefit from explicit cross-references in the text and consistent axis scaling to ease reader comparison.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the positive assessment of the dataset's potential impact and for the constructive feedback on annotation validation. We address the major comment below and will revise the manuscript to incorporate the requested quantitative analysis.

read point-by-point responses

  1. Referee: [Dataset construction / annotation interfaces] Section on dataset construction and crowd-sourcing interfaces (description of instance spotting and segmentation UIs): no inter-annotator agreement scores, no expert-vs-crowd IoU on a held-out subset, and no error analysis for boundary precision or category confusion are reported for the 2.5M instances. Because the central claim rests on these labels enabling precise localization in scene context, the absence of quantitative validation is load-bearing; systematic biases (e.g., under-segmentation of small or occluded objects) would directly undermine downstream utility.

    Authors: We agree that quantitative validation of the annotations is essential to support claims about precise localization in scene context. In the revised manuscript we will add a dedicated subsection to the dataset construction section that reports (1) inter-annotator agreement scores for both instance spotting and segmentation tasks, computed on a held-out set of images using multiple independent workers; (2) expert-versus-crowd IoU statistics on a separate held-out subset of 500 images annotated by computer-vision experts; and (3) a systematic error analysis that quantifies boundary-precision errors and category confusions, with explicit breakdowns for small and occluded objects. These additions will directly address concerns about potential systematic biases and strengthen the reproducibility of the benchmark. revision: yes

Circularity Check

0 steps flagged

No circularity: dataset construction and release paper with external benchmarks

full rationale

The paper presents the COCO dataset, its collection process via crowd-sourcing interfaces, statistical comparisons to PASCAL/ImageNet/SUN, and baseline results using an off-the-shelf Deformable Parts Model. No derivations, equations, predictions, or first-principles claims exist that could reduce to fitted inputs, self-definitions, or self-citation chains. The contribution is the independent dataset itself; any questions of label accuracy fall under correctness risk rather than circularity per the analysis rules. Steps array left empty as no load-bearing reductions were identified.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is a dataset construction paper with no mathematical derivations, fitted parameters, background axioms, or postulated entities beyond the empirical collection of images and labels.

pith-pipeline@v0.9.0 · 5480 in / 1095 out tokens · 42287 ms · 2026-05-11T23:56:56.623374+00:00 · methodology

0 comments
read the original abstract

We present a new dataset with the goal of advancing the state-of-the-art in object recognition by placing the question of object recognition in the context of the broader question of scene understanding. This is achieved by gathering images of complex everyday scenes containing common objects in their natural context. Objects are labeled using per-instance segmentations to aid in precise object localization. Our dataset contains photos of 91 objects types that would be easily recognizable by a 4 year old. With a total of 2.5 million labeled instances in 328k images, the creation of our dataset drew upon extensive crowd worker involvement via novel user interfaces for category detection, instance spotting and instance segmentation. We present a detailed statistical analysis of the dataset in comparison to PASCAL, ImageNet, and SUN. Finally, we provide baseline performance analysis for bounding box and segmentation detection results using a Deformable Parts Model.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 60 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Semantic Browsing: Controllable Diversity for Image Generation

    cs.CV 2026-06 unverdicted novelty 7.0

    A technique for controllable diversity in text-to-image generation by inducing structured semantic variations at the prompt level via VLM and agentic workflow.

  2. Beyond Linear Activation Steering: Invertible Latent Transformations for Controlling LLM Behavior

    cs.LG 2026-06 unverdicted novelty 7.0

    INNSteer learns an invertible neural network to map LLM activations into a latent space where linear steering becomes more effective, then applies the inverse map to produce nonlinear interventions in the original space.

  3. PlanBench-V: A Spatial Planning Map Benchmark for Vision-Language Models

    cs.CL 2026-06 unverdicted novelty 7.0

    PlanBench-V is a new benchmark and dataset for evaluating VLMs on spatial planning map interpretation via a four-stage framework of Perception, Reasoning, Association, and Implementation.

  4. FindIt: A Format-Informed Visual Detection Benchmark for Generalist Multimodal LLMs

    cs.CV 2026-06 unverdicted novelty 7.0

    FindIt is the first comprehensive benchmark for evaluating generalist MLLMs on promptable object detection, referring expression detection, instance-level detection, and video detection with standardized parsable outputs.

  5. Rethinking Visual Neglect: Steering via Context-Preference for MLLM Hallucination Mitigation

    cs.CL 2026-05 unverdicted novelty 7.0

    CAS mitigates object hallucinations in MLLMs by extracting two context preference vectors from designed conflict samples and applying signed residual injection at mid-early MLP layers without retraining or added latency.

  6. Structure over Pixels: Learning Variable-Length Visual Programs

    cs.CV 2026-05 unverdicted novelty 7.0

    STROP learns variable-length discrete visual programs for images by training a length head against frozen DINOv3 features in a four-phase curriculum while bypassing pixel reconstruction.

  7. Concept-Based Abductive and Contrastive Explanations for Behaviors of Vision Models

    cs.LG 2026-05 unverdicted novelty 7.0

    Concept-based abductive and contrastive explanations find minimal high-level concepts that causally determine vision model outcomes on individual images or groups sharing a specified behavior.

  8. Pro$^2$Assist: Continuous Step-Aware Proactive Assistance with Multimodal Egocentric Perception for Long-Horizon Procedural Tasks

    cs.AI 2026-05 unverdicted novelty 7.0

    Pro²Assist uses multimodal egocentric perception from AR glasses to track fine-grained progress in long-horizon procedural tasks and deliver timely proactive assistance, outperforming baselines by over 21% in action u...

  9. Beyond Accuracy: Benchmarking Cross-Task Consistency in Unified Multimodal Models

    cs.CV 2026-04 unverdicted novelty 7.0

    XTC-Bench reveals that strong performance on generation or understanding tasks in unified multimodal models does not guarantee cross-task semantic consistency, which instead depends on how tightly coupled the learning...

  10. Oracle Noise: Faster Semantic Spherical Alignment for Interpretable Latent Optimization

    cs.CV 2026-04 unverdicted novelty 7.0

    Oracle Noise optimizes diffusion model noise on a Riemannian hypersphere guided by key prompt words to preserve the Gaussian prior, eliminate norm inflation, and achieve faster semantic alignment than Euclidean methods.

  11. MMEB-V3: Measuring the Performance Gaps of Omni-Modality Embedding Models

    cs.IR 2026-04 unverdicted novelty 7.0

    MMEB-V3 benchmark shows omni-modality embedding models fail to enforce instruction-specified modality constraints and exhibit asymmetric, query-biased retrieval.

  12. Beyond Prompts: Unconditional 3D Inversion for Out-of-Distribution Shapes

    cs.CV 2026-04 unverdicted novelty 7.0

    Text-to-3D models lose prompt sensitivity for out-of-distribution shapes due to sink traps but retain geometric diversity via unconditional priors, enabling a decoupled inversion method for robust editing.

  13. Gaslight, Gatekeep, V1-V3: Early Visual Cortex Alignment Shields Vision-Language Models from Sycophantic Manipulation

    cs.CV 2026-04 conditional novelty 7.0

    Alignment of vision-language models with human V1-V3 early visual cortex negatively predicts resistance to sycophantic gaslighting attacks.

  14. MAST: Mask-Guided Attention Mass Allocation for Training-Free Multi-Style Transfer

    cs.CV 2026-04 unverdicted novelty 7.0

    MAST is a mask-guided attention allocation method that enables artifact-free multi-style transfer in diffusion models by anchoring layout, distributing attention mass, scaling sharpness, and injecting details.

  15. WildDet3D: Scaling Promptable 3D Detection in the Wild

    cs.CV 2026-04 unverdicted novelty 7.0

    WildDet3D is a promptable 3D detector paired with a new 1M-image dataset across 13.5K categories that sets SOTA on open-world and zero-shot 3D detection benchmarks.

  16. Descriptor: Parasitoid Wasps and Associated Hymenoptera Dataset (DAPWH)

    cs.CV 2026-02 unverdicted novelty 7.0

    Releases the DAPWH dataset of 3556 wasp images including 1739 COCO-annotated examples to enable AI models for identifying Ichneumonoidea and associated families.

  17. Event-based Civil Infrastructure Visual Defect Detection: ev-CIVIL Dataset and Benchmark

    cs.CV 2025-04 unverdicted novelty 7.0

    Presents the ev-CIVIL dataset and benchmark showing that event-based cameras can support real-time detection of cracks and spalling in civil infrastructure under challenging lighting.

  18. PromptGuard: Soft Prompt-Guided Unsafe Content Moderation for Text-to-Image Models

    cs.CV 2025-01 unverdicted novelty 7.0

    PromptGuard optimizes a universal safety soft prompt (and category-specific variants) in T2I embedding space to moderate NSFW inputs, achieving average unsafe ratios of 5.84-6.18% while being 3.8x faster than prior defenses.

  19. Hierarchical Text-Conditional Image Generation with CLIP Latents

    cs.CV 2022-04 accept novelty 7.0

    A hierarchical prior-decoder model using CLIP latents generates more diverse text-conditional images than direct methods while preserving photorealism and caption fidelity.

  20. High-Resolution Image Synthesis with Latent Diffusion Models

    cs.CV 2021-12 conditional novelty 7.0

    Latent diffusion models achieve state-of-the-art inpainting and competitive results on unconditional generation, scene synthesis, and super-resolution by performing the diffusion process in the latent space of pretrai...

  21. U4D: Unsupervised 4D Dynamic Scene Understanding

    cs.CV 2019-07 unverdicted novelty 7.0

    Unsupervised joint semantic instance segmentation, 4D reconstruction, and scene flow from multi-view video of multi-person dynamic scenes, with reported ~40% gains over prior methods.

  22. Concept Removal for Frontier Image Generative Models

    cs.CV 2026-06 unverdicted novelty 6.0

    A transcoder-based in-place replacement of the bottleneck layer enables selective concept removal in modern diffusion and autoregressive image models without degrading output quality.

  23. HANCLIP: A Family of Hyperbolic Angular Negation Vision Language Models

    cs.CV 2026-06 unverdicted novelty 6.0

    HANCLIP restructures VLM embeddings with hyperbolic space and angular negation objectives to raise negation sensitivity on NegBench while keeping standard retrieval and classification performance.

  24. Unmasking LAION-5B: Age, Gender, Race, and Emotion Biases in Large-Scale Image Datasets

    cs.CV 2026-06 unverdicted novelty 6.0

    Empirical audit of LAION-2B-en and LAION-2B-multi finds overrepresentation of young adults, White people, and males plus stereotypical emotion associations across two attribute classifiers.

  25. Time-frequency localization of bird calls in dense soundscapes

    cs.SD 2026-06 unverdicted novelty 6.0

    YOLO11 models localize bird vocalizations on spectrograms, nearly doubling IoMin@50 F1 on Singapore data and outperforming baselines on Hawaii recordings.

  26. Maximum Matching Accuracy: An Instance Segmentation Evaluation Metric Utilizing Globally Optimal Matching

    cs.CV 2026-06 unverdicted novelty 6.0

    MMA is a threshold-free continuous metric for instance segmentation that uses globally optimal bipartite matching between predictions and ground truth followed by per-pixel normalization to aggregate overlap.

  27. Beyond Compression: Quantifying Spectral Accessibility in Vision Representations

    cs.CV 2026-06 unverdicted novelty 6.0

    Vision encoders alter spectral accessibility non-monotonically across depth with architecture-specific effects from projections and pooling, quantified via a new residual loss against random baselines.

  28. VLA-Trace: Diagnosing Vision-Language-Action Models through Representation and Behavior Tracing

    cs.AI 2026-05 unverdicted novelty 6.0

    VLA-Trace diagnoses two VLA models via representation CKA, attention interventions, and behavioral tests, finding distinct finetuning dynamics, different routing, and strong visual grounding but weak fine-grained sema...

  29. Trajectory-Consistent Calibration for Cache-Accelerated Diffusion Models

    cs.CV 2026-05 unverdicted novelty 6.0

    TCC calibrates cached representations in diffusion sampling via an offline iterative procedure that accounts for trajectory shifts, improving FID from 29.83 to 27.35 on PixArt-alpha while preserving reuse policies.

  30. Why We Look Where We Look: Emergent Human-like Fixations of a Foveated Visual Language Model Maximizing Scene Understanding

    cs.CV 2026-05 unverdicted novelty 6.0

    A foveated VLM trained for scene comprehension produces human-like fixations, outperforming models trained for search, classification, or with altered peripheral vision.

  31. SparseSAM: Structured Sparsification of Activations in Segment Anything Models

    cs.CV 2026-05 unverdicted novelty 6.0

    SparseSAM achieves 2x faster inference and 2.8x memory reduction in SAM with only 0.004 mIoU loss at 0.4 density via Stripe-Sort Attention and Residual-Consistency MLP.

  32. CASCADE: Context-Aware Relaxation for Speculative Image Decoding

    cs.CV 2026-05 unverdicted novelty 6.0

    CASCADE formalizes semantic interchangeability and convergence in target model representations to enable context-aware acceptance relaxation in tree-based speculative decoding, delivering up to 3.6x speedup on text-to...

  33. Semantics-Aware Hierarchical Token Communication: Clustering, Bit Mapping, and Power Allocation

    eess.SP 2026-04 unverdicted novelty 6.0

    H-TokCom groups tokens by semantic similarity and protects cluster-level bits with higher power, raising semantic similarity from 0.206 to 0.279 at 3 dB SNR on COCO data.

  34. Delta Score Matters! Spatial Adaptive Multi Guidance in Diffusion Models

    cs.CV 2026-04 unverdicted novelty 6.0

    SAMG uses spatially adaptive guidance scales derived from a geometric analysis of classifier-free guidance to resolve the detail-artifact dilemma in diffusion-based image and video generation.

  35. LatentDiff: Scaling Semantic Dataset Comparison to Millions of Images

    cs.CV 2026-04 unverdicted novelty 6.0

    LatentDiff scales semantic dataset comparison to millions of images using latent spaces of vision encoders combined with sparse autoencoders and density ratio estimation, showing better accuracy and robustness than ca...

  36. DeepSignature: Digitally Signed, Content-Encoding Watermarks for Robust and Transparent Image Authentication

    cs.CR 2026-04 unverdicted novelty 6.0

    DeepSignature embeds digitally signed content-encoding watermarks via neural networks for robust image authentication, source attribution, and latent-space tamper localization.

  37. Source-Modality Monitoring in Vision-Language Models

    cs.CL 2026-04 unverdicted novelty 6.0

    Vision-language models use semantic signals more than syntactic ones to bind words like 'image' to actual visual inputs, with implications for robustness in multimodal systems.

  38. MACS: Modality-Aware Capacity Scaling for Efficient Multimodal MoE Inference

    cs.LG 2026-04 unverdicted novelty 6.0

    MACS improves MoE MLLM inference efficiency via entropy-weighted token loads and dynamic modality-adaptive expert capacity allocation.

  39. MACS: Modality-Aware Capacity Scaling for Efficient Multimodal MoE Inference

    cs.LG 2026-04 unverdicted novelty 6.0

    MACS improves inference speed in multimodal MoE models by entropy-weighted balancing of visual tokens and real-time modality-adaptive expert capacity allocation.

  40. Zero-shot World Models Are Developmentally Efficient Learners

    cs.AI 2026-04 unverdicted novelty 6.0

    A zero-shot visual world model trained on one child's experience achieves broad competence on physical understanding benchmarks while matching developmental behavioral patterns.

  41. Bayesian Optimization for Mixed-Variable Problems in the Natural Sciences

    cs.LG 2026-04 unverdicted novelty 6.0

    A generalization of probabilistic reparameterization allows gradient-based acquisition optimization in fully mixed-variable Bayesian optimization with Gaussian process surrogates for non-equidistant discrete spaces.

  42. DeCo-DETR: Decoupled Cognition DETR for efficient Open-Vocabulary Object Detection

    cs.CV 2026-04 unverdicted novelty 6.0

    DeCo-DETR builds hierarchical semantic prototypes offline and uses decoupled training streams to deliver competitive zero-shot open-vocabulary detection with improved inference speed.

  43. Chasing Ghosts: A Simulation-to-Real Olfactory Navigation Stack with Optional Vision Augmentation

    cs.RO 2026-02 unverdicted novelty 6.0

    A simulation-to-real navigation policy enables a quadrotor to locate an odor source using only basic olfaction sensors and optional vision, validated in indoor real-world flights.

  44. Adaptive Residual-Update Steering for Low-Overhead Hallucination Mitigation in Large Vision Language Models

    cs.CV 2025-11 unverdicted novelty 6.0

    RUDDER creates a persistent visual anchor by extracting CARD from prefill residuals and modulating its injection via an adaptive Beta Gate, cutting CHAIR_S by 24.4% and CHAIR_i by 23.6% on average across LLaVA, Idefic...

  45. Routing-Based Continual Learning for Multimodal Large Language Models

    cs.LG 2025-11 unverdicted novelty 6.0

    Routing architecture for MLLMs enables continual learning with constant compute, matching multi-task learning performance and supporting cross-modal transfer.

  46. Navigating the Challenges of AI-Generated Image Detection in the Wild: What Truly Matters?

    cs.CV 2025-07 conditional novelty 6.0

    The ITW-SM dataset and targeted optimization of detector design choices yield a 26.87% average AUC improvement for state-of-the-art AI-generated image detectors under real-world social media conditions.

  47. Expressive yet Efficient Feature Expansion with Adaptive Cross-Hadamard Products

    cs.CV 2025-05 unverdicted novelty 6.0

    Proposes ACH module with differentiable sampling and softsign normalization for efficient feature expansion, integrated via NAS into Hadaptive-Net to claim SOTA accuracy/speed trade-offs on image classification.

  48. MM1: Methods, Analysis & Insights from Multimodal LLM Pre-training

    cs.CV 2024-03 unverdicted novelty 6.0

    MM1 models achieve state-of-the-art few-shot multimodal results by pre-training on a careful mix of image-caption, interleaved, and text-only data with optimized image encoders.

  49. Informative Image Captioning with External Sources of Information

    cs.CL 2019-06 unverdicted novelty 6.0

    A multimodal Transformer ingests image features plus multiple external entity label sources and learns to control their appearance in fluent output captions.

  50. On Physical Adversarial Patches for Object Detection

    cs.CV 2019-06 unverdicted novelty 6.0

    A physical patch suppresses all object detections by YOLOv3 even for distant objects without overlapping them.

  51. TuringViT: Making SOTA Vision Transformers Accessible to All

    cs.CV 2026-06 unverdicted novelty 5.0

    TuringViT claims a new ViT design with linear attention and curated data that matches SOTA performance using 10% of typical pretraining data while supporting dynamic resolutions and improving VLM integration.

  52. TuringViT: Making SOTA Vision Transformers Accessible to All

    cs.CV 2026-06 unverdicted novelty 5.0

    TuringViT uses Turing Linear Attention, VISTA-Curation, and dynamic-resolution pretraining to outperform open ViT baselines with 10% data while improving VLM performance and high-resolution latency.

  53. Accelerating Vision Foundation Models with Drop-in Depthwise Convolution

    cs.CV 2026-05 unverdicted novelty 5.0

    Replacing selected attention heads in pretrained ViTs with depthwise convolutions, identified by simple strategies and recovered via fine-tuning, delivers 17-20% inference speedup on image tasks with minimal accuracy loss.

  54. Replacement Learning: Training Neural Networks with Fewer Parameters

    cs.CV 2026-05 unverdicted novelty 5.0

    Replacement Learning replaces selected blocks in CNNs and ViTs with learnable parameter-fusion surrogates derived from adjacent layers to reduce full-depth backpropagation redundancy.

  55. KVCapsule: Efficient Sequential KV Cache Compression for Vision-Language Models with Asymmetric Redundancy

    cs.CV 2026-05 unverdicted novelty 5.0

    KVCapsule compresses KV cache in VLMs by 60% to deliver up to 2x higher tokens-per-second and 2.4x memory reduction with negligible accuracy loss.

  56. Global Context or Local Detail? Adaptive Visual Grounding for Hallucination Mitigation

    cs.CV 2026-04 unverdicted novelty 5.0

    PND mitigates object hallucination in vision-language models via dual-path contrastive decoding that boosts visual evidence and penalizes linguistic priors, yielding up to 6.5% gains on POPE, MME, and CHAIR benchmarks.

  57. From Codebooks to VLMs: Evaluating Automated Visual Discourse Analysis for Climate Change on Social Media

    cs.CV 2026-04 unverdicted novelty 5.0

    VLMs recover reliable population-level trends in climate change visual discourse on social media even when per-image accuracy is only moderate.

  58. FineEdit: Fine-Grained Image Edit with Bounding Box Guidance

    cs.CV 2026-04 unverdicted novelty 5.0

    FineEdit adds multi-level bounding box injection to diffusion image editing, releases a 1.2M-pair dataset with box annotations, and shows better instruction following and background consistency than prior open models ...

  59. DeCo-DETR: Decoupled Cognition DETR for efficient Open-Vocabulary Object Detection

    cs.CV 2026-04 unverdicted novelty 5.0

    DeCo-DETR constructs a hierarchical semantic prototype space from LVLM-generated descriptions aligned via CLIP and uses decoupled training streams to separate semantic reasoning from detection, yielding efficient open...

  60. Language-Pretraining-Induced Bias: A Strong Foundation for General Vision Tasks

    cs.CV 2026-04 unverdicted novelty 5.0

    Random label bridge training aligns LLM parameters with vision tasks, and partial training of certain layers often suffices due to their foundational properties.

Reference graph

Works this paper leans on

51 extracted references · 51 canonical work pages · cited by 76 Pith papers

  1. [1]

    Im- ageNet: A Large-Scale Hierarchical Image Database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Im- ageNet: A Large-Scale Hierarchical Image Database,” in CVPR, 2009

    work page 2009

  2. [2]

    The PASCAL visual object classes (VOC) challenge,

    M. Everingham, L. Van Gool, C. K. I. Williams, J. Winn, and A. Zis- serman, “The PASCAL visual object classes (VOC) challenge,” IJCV, vol. 88, no. 2, pp. 303–338, Jun. 2010

    work page 2010

  3. [3]

    SUN database: Large-scale scene recognition from abbey to zoo,

    J. Xiao, J. Hays, K. A. Ehinger, A. Oliva, and A. Torralba, “SUN database: Large-scale scene recognition from abbey to zoo,” in CVPR, 2010

    work page 2010

  4. [4]

    Pedestrian detec- tion: An evaluation of the state of the art,

    P . Doll ´ar, C. Wojek, B. Schiele, and P . Perona, “Pedestrian detec- tion: An evaluation of the state of the art,” P AMI, vol. 34, 2012

    work page 2012

  5. [5]

    ImageNet classifica- tion with deep convolutional neural networks,

    A. Krizhevsky, I. Sutskever, and G. Hinton, “ImageNet classifica- tion with deep convolutional neural networks,” in NIPS, 2012

    work page 2012

  6. [6]

    Rich feature hierarchies for accurate object detection and semantic segmenta- tion,

    R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature hierarchies for accurate object detection and semantic segmenta- tion,” in CVPR, 2014

    work page 2014

  7. [7]

    OverFeat: Integrated recognition, localization and detection using convolutional networks,

    P . Sermanet, D. Eigen, S. Zhang, M. Mathieu, R. Fergus, and Y. LeCun, “OverFeat: Integrated recognition, localization and detection using convolutional networks,” in ICLR, April 2014

    work page 2014

  8. [8]

    Describing objects by their attributes,

    A. Farhadi, I. Endres, D. Hoiem, and D. Forsyth, “Describing objects by their attributes,” in CVPR, 2009

    work page 2009

  9. [9]

    SUN attribute database: Discovering, annotating, and recognizing scene attributes,

    G. Patterson and J. Hays, “SUN attribute database: Discovering, annotating, and recognizing scene attributes,” in CVPR, 2012

    work page 2012

  10. [10]

    Poselets: Body part detectors trained using 3D human pose annotations,

    L. Bourdev and J. Malik, “Poselets: Body part detectors trained using 3D human pose annotations,” in ICCV, 2009

    work page 2009

  11. [11]

    Indoor seg- mentation and support inference from RGBD images,

    N. Silberman, D. Hoiem, P . Kohli, and R. Fergus, “Indoor seg- mentation and support inference from RGBD images,” in ECCV, 2012

    work page 2012

  12. [12]

    Canonical perspective and the perception of objects,

    S. Palmer, E. Rosch, and P . Chase, “Canonical perspective and the perception of objects,” Attention and performance IX , vol. 1, p. 4, 1981

    work page 1981

  13. [13]

    Diagnosing error in object detectors,

    Hoiem, D, Y. Chodpathumwan, and Q. Dai, “Diagnosing error in object detectors,” in ECCV, 2012

    work page 2012

  14. [14]

    Semantic object classes in video: A high-definition ground truth database,

    G. Brostow, J. Fauqueur, and R. Cipolla, “Semantic object classes in video: A high-definition ground truth database,” PRL, vol. 30, no. 2, pp. 88–97, 2009

    work page 2009

  15. [15]

    LabelMe: a database and web-based tool for image annotation,

    B. Russell, A. Torralba, K. Murphy, and W. Freeman, “LabelMe: a database and web-based tool for image annotation,” IJCV, vol. 77, no. 1-3, pp. 157–173, 2008

    work page 2008

  16. [16]

    OpenSurfaces: A richly annotated catalog of surface appearance,

    S. Bell, P . Upchurch, N. Snavely, and K. Bala, “OpenSurfaces: A richly annotated catalog of surface appearance,” SIGGRAPH, vol. 32, no. 4, 2013

    work page 2013

  17. [17]

    Im2text: Describing images using 1 million captioned photographs

    V . Ordonez, G. Kulkarni, and T. Berg, “Im2text: Describing images using 1 million captioned photographs.” in NIPS, 2011

    work page 2011

  18. [18]

    Scalable multi-label annotation,

    J. Deng, O. Russakovsky, J. Krause, M. Bernstein, A. Berg, and L. Fei-Fei, “Scalable multi-label annotation,” in CHI, 2014

    work page 2014

  19. [19]

    Microsoft COCO: Common objects in context,

    T. Lin, M. Maire, S. Belongie, J. Hays, P . Perona, D. Ramanan, P . Doll´ar, and C. L. Zitnick, “Microsoft COCO: Common objects in context,” in ECCV, 2014

    work page 2014

  20. [20]

    A taxonomy and evaluation of dense two-frame stereo correspondence algorithms,

    D. Scharstein and R. Szeliski, “A taxonomy and evaluation of dense two-frame stereo correspondence algorithms,” IJCV, vol. 47, no. 1-3, pp. 7–42, 2002

    work page 2002

  21. [21]

    A database and evaluation methodology for optical flow,

    S. Baker, D. Scharstein, J. Lewis, S. Roth, M. Black, and R. Szeliski, “A database and evaluation methodology for optical flow,” IJCV, vol. 92, no. 1, pp. 1–31, 2011

    work page 2011

  22. [22]

    Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories,

    L. Fei-Fei, R. Fergus, and P . Perona, “Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories,” in CVPR Workshop of Generative Model Based Vision (WGMBV) , 2004

    work page 2004

  23. [23]

    Caltech-256 object category dataset,

    G. Griffin, A. Holub, and P . Perona, “Caltech-256 object category dataset,” California Institute of Technology, Tech. Rep. 7694, 2007

    work page 2007

  24. [24]

    Histograms of oriented gradients for human detection,

    N. Dalal and B. Triggs, “Histograms of oriented gradients for human detection,” in CVPR, 2005

    work page 2005

  25. [25]

    The MNIST database of handwritten digits,

    Y. Lecun and C. Cortes, “The MNIST database of handwritten digits,” 1998. [Online]. Available: http://yann.lecun.com/exdb/ mnist/

    work page 1998

  26. [26]

    Columbia object image library (coil-20),

    S. A. Nene, S. K. Nayar, and H. Murase, “Columbia object image library (coil-20),” Columbia Universty, Tech. Rep., 1996

    work page 1996

  27. [27]

    Learning multiple layers of fea- tures from tiny images,

    A. Krizhevsky and G. Hinton, “Learning multiple layers of fea- tures from tiny images,” Computer Science Department, University of T oronto, T ech. Rep, 2009

    work page 2009

  28. [28]

    80 million tiny images: A large data set for nonparametric object and scene recognition,

    A. Torralba, R. Fergus, and W. T. Freeman, “80 million tiny images: A large data set for nonparametric object and scene recognition,” P AMI, vol. 30, no. 11, pp. 1958–1970, 2008

    work page 1958

  29. [29]

    From large scale image categorization to entry-level categories,

    V . Ordonez, J. Deng, Y. Choi, A. Berg, and T. Berg, “From large scale image categorization to entry-level categories,” in ICCV, 2013

    work page 2013

  30. [30]

    Fellbaum, WordNet: An electronic lexical database

    C. Fellbaum, WordNet: An electronic lexical database . Blackwell Books, 1998

    work page 1998

  31. [31]

    Caltech-UCSD Birds 200,

    P . Welinder, S. Branson, T. Mita, C. Wah, F. Schroff, S. Belongie, and P . Perona, “Caltech-UCSD Birds 200,” Caltech, Tech. Rep. CNS-TR-201, 2010

    work page 2010

  32. [32]

    Face detection: A survey,

    E. Hjelm ˚as and B. Low, “Face detection: A survey,” CVIU, vol. 83, no. 3, pp. 236–274, 2001

    work page 2001

  33. [33]

    Labeled faces in the wild,

    G. B. Huang, M. Ramesh, T. Berg, and E. Learned-Miller, “Labeled faces in the wild,” University of Massachusetts, Amherst, Tech. Rep. 07-49, October 2007

    work page 2007

  34. [34]

    Detecting avocados to zucchinis: what have we done, and where are we going?

    O. Russakovsky, J. Deng, Z. Huang, A. Berg, and L. Fei-Fei, “Detecting avocados to zucchinis: what have we done, and where are we going?” in ICCV, 2013

    work page 2013

  35. [35]

    TextonBoost for image understanding: Multi-class object recognition and seg- mentation by jointly modeling texture, layout, and context,

    J. Shotton, J. Winn, C. Rother, and A. Criminisi, “TextonBoost for image understanding: Multi-class object recognition and seg- mentation by jointly modeling texture, layout, and context,” IJCV, vol. 81, no. 1, pp. 2–23, 2009

    work page 2009

  36. [36]

    A comparison and evaluation of multi-view stereo reconstruction algorithms,

    S. M. Seitz, B. Curless, J. Diebel, D. Scharstein, and R. Szeliski, “A comparison and evaluation of multi-view stereo reconstruction algorithms,” in CVPR, 2006

    work page 2006

  37. [37]

    Contour detec- tion and hierarchical image segmentation,

    P . Arbelaez, M. Maire, C. Fowlkes, and J. Malik, “Contour detec- tion and hierarchical image segmentation,” P AMI, vol. 33, no. 5, pp. 898–916, 2011

    work page 2011

  38. [38]

    Learning to detect unseen object classes by between-class attribute transfer,

    C. Lampert, H. Nickisch, and S. Harmeling, “Learning to detect unseen object classes by between-class attribute transfer,” in CVPR, 2009

    work page 2009

  39. [39]

    Learning spatial context: Using stuff to find things,

    G. Heitz and D. Koller, “Learning spatial context: Using stuff to find things,” in ECCV, 2008

    work page 2008

  40. [40]

    Sitton, Spelling Sourcebook

    R. Sitton, Spelling Sourcebook . Egger Publishing, 1996

    work page 1996

  41. [41]

    Finding iconic images,

    T. Berg and A. Berg, “Finding iconic images,” in CVPR, 2009

    work page 2009

  42. [42]

    Unbiased look at dataset bias,

    A. Torralba and A. Efros, “Unbiased look at dataset bias,” in CVPR, 2011

    work page 2011

  43. [43]

    Evaluation of gist descriptors for web-scale image search,

    M. Douze, H. J ´egou, H. Sandhawalia, L. Amsaleg, and C. Schmid, “Evaluation of gist descriptors for web-scale image search,” in CIVR, 2009

    work page 2009

  44. [44]

    Object detection with discriminatively trained part-based mod- els,

    P . Felzenszwalb, R. Girshick, D. McAllester, and D. Ramanan, “Object detection with discriminatively trained part-based mod- els,” P AMI, vol. 32, no. 9, pp. 1627–1645, 2010

    work page 2010

  45. [45]

    Discriminatively trained deformable part models, release 5,

    R. Girshick, P . Felzenszwalb, and D. McAllester, “Discriminatively trained deformable part models, release 5,” P AMI, 2012

    work page 2012

  46. [46]

    Do we need more training data or better models for object detection?

    X. Zhu, C. Vondrick, D. Ramanan, and C. Fowlkes, “Do we need more training data or better models for object detection?” in BMVC, 2012

    work page 2012

  47. [47]

    Object segmentation by alignment of poselet activations to image contours,

    T. Brox, L. Bourdev, S. Maji, and J. Malik, “Object segmentation by alignment of poselet activations to image contours,” in CVPR, 2011

    work page 2011

  48. [48]

    Layered object models for image segmentation,

    Y. Yang, S. Hallman, D. Ramanan, and C. Fowlkes, “Layered object models for image segmentation,” P AMI, vol. 34, no. 9, pp. 1731–1743, 2012

    work page 2012

  49. [49]

    Using segmentation to verify object hypotheses,

    D. Ramanan, “Using segmentation to verify object hypotheses,” in CVPR, 2007

    work page 2007

  50. [50]

    Learning to localize detected objects,

    Q. Dai and D. Hoiem, “Learning to localize detected objects,” in CVPR, 2012

    work page 2012

  51. [51]

    Col- lecting image annotations using Amazon’s Mechanical Turk,

    C. Rashtchian, P . Young, M. Hodosh, and J. Hockenmaier, “Col- lecting image annotations using Amazon’s Mechanical Turk,” in NAACL Workshop , 2010. 13 Fig. 11: Icons of 91 categories in the MS COCO dataset grouped by 11 super-categories. We use these icons in our annotation pipeline to help workers quickly reference the indicated object category. Fig. 12:...

    work page 2010