Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2205.04596

When does dough become a bagel? Analyzing the remaining mistakes on ImageNet

Vijay Vasudevan·Benjamin Caine·Raphael Gontijo-Lopes·Sara Fridovich-Keil·Rebecca Roelofs
View original source ↗Read source document ↗Export metadata ↓

Abstract

Image classification accuracy on the ImageNet dataset has been a barometer for progress in computer vision over the last decade. Several recent papers have questioned the degree to which the benchmark remains useful to the community, yet innovations continue to contribute gains to performance, with today's largest models achieving 90%+ top-1 accuracy. To help contextualize progress on ImageNet and provide a more meaningful evaluation for today's state-of-the-art models, we manually review and categorize every remaining mistake that a few top models make in order to provide insight into the long-tail of errors on one of the most benchmarked datasets in computer vision. We focus on the multi-label subset evaluation of ImageNet, where today's best models achieve upwards of 97% top-1 accuracy. Our analysis reveals that nearly half of the supposed mistakes are not mistakes at all, and we uncover new valid multi-labels, demonstrating that, without careful review, we are significantly underestimating the performance of these models. On the other hand, we also find that today's best models still make a significant number of mistakes (40%) that are obviously wrong to human reviewers. To calibrate future progress on ImageNet, we provide an updated multi-label evaluation set, and we curate ImageNet-Major: a 68-example "major error" slice of the obvious mistakes made by today's top models -- a slice where models should achieve near perfection, but today are far from doing so.

Explore related subjects

cs.CV

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Vijay Vasudevan, Benjamin Caine, Raphael Gontijo-Lopes, Sara Fridovich-Keil, Rebecca Roelofs. 2022-05-25. When does dough become a bagel? Analyzing the remaining mistakes on ImageNet. https://arxiv.org/abs/2205.04596

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

SAMReg: SAM-enabled Image Registration with ROI-based Correspondence

This paper describes a new spatial correspondence representation based on paired regions-of-interest (ROIs), for medical image registration. The distinct properties of the proposed ROI-based correspondence are discussed, in the context of potential benefits in clinical applications following image registration, compared with alternative correspondence-representing approaches, such as those based on sampled displacements and spatial transformation functions. These benefits include a clear connection between learning-based image registration and segmentation, which in turn motivates two cases of image registration approaches using (pre-)trained segmentation networks. Based on the segment anything model (SAM), a vision foundation model for segmentation, we develop a new registration algorithm SAMReg, which does not require any training (or training data), gradient-based fine-tuning or prompt engineering. The proposed SAMReg models are evaluated across five real-world applications, including intra-subject registration tasks with cardiac MR and lung CT, challenging inter-subject registration scenarios with prostate MR and retinal imaging, and an additional evaluation with a non-clinical example with aerial image registration. The proposed methods outperform both intensity-based iterative algorithms and DDF-predicting learning-based networks across tested metrics including Dice and target registration errors on anatomical structures, and further demonstrates competitive performance compared to weakly-supervised registration approaches that rely on fully-segmented training data. Open source code and examples are available at: https://github.com/sqhuang0103/SAMReg.git.

cs.CV↗
arXiv2026

Q-SiT: Teaching LMMs for Image Quality Scoring and Interpreting

Image quality scoring and interpreting are two fundamental components of Image Quality Assessment (IQA). The former quantifies image quality, while the latter enables descriptive question answering about image quality. Traditionally, these two tasks have been addressed independently. However, image-quality-specific psychophysical studies suggest that these two tasks are conceptually interconnected: interpreting explicitly represents perceived quality attributes whereas scoring summarizes such evidence into an overall quality judgment. Thus, unifying these capabilities within a single model is both intuitive and logically coherent. In this paper, we propose Q-SiT (Quality Scoring and Interpreting joint Teaching), a unified framework that enables large multimodal models (LMMs) to learn both image quality scoring and interpreting simultaneously. We achieve this by transforming conventional IQA datasets into learnable question-answering datasets and incorporating human-annotated quality interpreting data for training. Furthermore, we introduce an efficient scoring \& interpreting balance strategy, which first determines the optimal data mix ratio on lightweight LMMs and then maps this ratio to primary LMMs for fine-tuning adjustment. This strategy not only mitigates task interference and enhances cross-task knowledge transfer but also significantly reduces computational costs compared to direct optimization on full-scale LMMs. With this joint learning framework and corresponding training strategy, we develop Q-SiT, the first model capable of simultaneously performing image quality scoring and interpreting tasks, along with its lightweight variant, Q-SiT-mini. Experimental results demonstrate that Q-SiT achieves strong performance in both tasks with superior generalization IQA abilities, while Q-SiT-mini significantly reduces computational overhead while maintaining competitive performance.

cs.CV↗
arXiv2026

Preserving Guidance in Cost-Volume Retrieval under Extreme LiDAR Sparsity in Iterative Stereo

While accurate LiDAR depth has been shown to improve stereo matching, high-end LiDAR remains costly and difficult to deploy at scale, motivating guidance from sparse LiDAR measurements. In this paper, we revisit how extremely sparse LiDAR can guide iterative stereo by examining its cost-volume retrieval mechanism. Our analysis shows that LiDAR guidance degrades sharply when only a few hundred points are available, and we provide a signal-processing explanation for why iterative stereo fails to fully exploit such sparse input. This insight leads to a simple yet highly effective remedy: pre-filling the initial disparity map so that cost-volume retrieval becomes more reliable under extremely sparse LiDAR. We find that pre-filling is also effective when injecting LiDAR depth into image features via early fusion, but for a fundamentally different reason, necessitating a distinct pre-filling approach. Combining both forms of guidance, our proposed Guided RAFT-Stereo (GRAFT-Stereo) achieves substantial improvements over existing LiDAR-guided stereo methods across multiple datasets.

cs.CV↗