# Image Classification on ImageNet-100 (test) Robustness Source: [Image Classification on ImageNet-100 (test) Robustness]() Collection index: [llms.txt]() Data updated: 2026-06-25T08:46:53.509Z Evaluation of Image Classification robustness on the test split of the ImageNet-100 dataset, specifically measuring Top-1 Accuracy under various visual corruptions (Noise, Blur, Brightness, Contrast, Flips). A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Image Classification | | Dataset (including any stated split) | ImageNet-100 (test) | | Metric | Top-1 Accuracy (Original) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 4 of 4 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Top-1 Accuracy (Original)](): Higher is better; unit: PERCENT. - [Top-1 Accuracy (Gaussian Noise)](): Higher is better; unit: PERCENT. - [Top-1 Accuracy (Gaussian Blur)](): Higher is better; unit: PERCENT. - [Top-1 Accuracy (Brightness)](): Higher is better; unit: PERCENT. - [Top-1 Accuracy (Contrast)](): Higher is better; unit: PERCENT. - [Top-1 Accuracy (Vertical Flip)](): Higher is better; unit: PERCENT. - [Top-1 Accuracy (Horizontal Flip)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Top-1 Accuracy (Original) | Result date | Report paper | | --- | --- | --- | --- | | SS-CA | 91.14 | Not recorded | 691bfdf36750be83fe5ceb57 | | Chen et al. | 89.83 | Not recorded | 691bfdf36750be83fe5ceb57 | | Xiao et al. | 89.77 | Not recorded | 691bfdf36750be83fe5ceb57 | | Conventional Training | 89.5 | Not recorded | 691bfdf36750be83fe5ceb57 | ### SS-CA Top-1 Accuracy (Original): 91.14. Result ID: 6a3ceaeafbc373592c52a5e7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"CLIP (ViT /32b)"}. This record does not assert independent reproduction. ### Chen et al. Top-1 Accuracy (Original): 89.83. Result ID: 6a3ceaeafbc373592c52a5e6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"CLIP (ViT /32b)"}. This record does not assert independent reproduction. ### Xiao et al. Top-1 Accuracy (Original): 89.77. Result ID: 6a3ceaeafbc373592c52a5e5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"CLIP (ViT /32b)"}. This record does not assert independent reproduction. ### Conventional Training Top-1 Accuracy (Original): 89.5. Result ID: 6a3ceaeafbc373592c52a5e4. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"CLIP (ViT /32b)"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.