Newest papers

Revealing Hidden Failure Modes in Chest X-ray Classification via Spectral Domain Analysis

Samuel Halimi, Loic Themyr, Arnaud Abreu · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.29830 · cited by 0 works (OpenAlex) · open-access PDF

Indonesian

Model pembelajaran mendalam untuk deteksi anomali X-ray dada tetap rentan terhadap pergeseran distribusi halus (misalnya, teknik akuisisi, faktor terkait pasien, dan preprocessing). Analisis kesalahan tradisional sering bergantung pada metadata semantik atau embeddings model, yang dapat menyamarkan variasi sinyal tingkat rendah yang merusak kinerja. Dalam pekerjaan ini, kami mengusulkan kerangka kerja berpusat data untuk penemuan mode kegagalan otomatis menggunakan analisis spektral. Kami memancarkan gambar ke dalam domain frekuensi dan mengekstrak profil kompak yang merangkum distribusi energi sinyal di seluruh band frekuensi. Dengan melakukan clustering yang tidak diawasi pada profil spektral ini, kami menunjukkan bahwa kegagalan model tidak didistribusikan secara acak, tetapi sangat terkonsentrasi dalam cluster spektral tertentu. Metode ini secara efektif meng

Machine translation (Meta M2M100); the English original follows.

Abstract

Deep learning models for chest X-ray anomaly detection remain vulnerable to subtle distributional shifts (e.g., acquisition technique, patient-related factors, and preprocessing). Traditional error analysis often relies on semantic metadata or model embeddings, which can mask low-level signal variations that degrade performance. In this work, we propose a data-centric framework for automated failure mode discovery using spectral analysis. We project images into the frequency domain and extract a compact profile summarizing the distribution of signal energy across frequency bands. By performing unsupervised clustering on these spectral profiles, we demonstrate that model failures are not randomly distributed, but are strongly concentrated within specific spectral clusters. This method effectively isolates "blind spots", enabling the prediction of model reliability and the discovery of performance-degrading data slices without requiring ground-truth failure annotations.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Artificial intelligence, Reliability (semiconductor), Pattern recognition (psychology), Cluster analysis, Spectral clustering

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange