{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T19:09:36Z","timestamp":1784056176787,"version":"3.55.0"},"reference-count":0,"publisher":"Centro Latino Americano de Estudios en Informatica","issue":"3","license":[{"start":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T00:00:00Z","timestamp":1783987200000},"content-version":"unspecified","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["CLEIej"],"abstract":"<jats:p>Technology played a crucial role in combating the COVID pandemic, both in the rapid\r\ndevelopment of vaccines and the early detection of the virus. Consequently, numerous\r\nstudies in the medical field have focused on leveraging the power of artificial intelligence\r\nfor COVID-19 detection. However, in the medical domain, it is essential to have a clear\r\nunderstanding of the processes and algorithms used in decision-making, as these directly\r\nimpact people\u2019s health. Therefore, efforts have been made to implement explainable\r\nartificial intelligence techniques, enabling humans to understand and explain the deep\r\nlearning algorithms used in disease detection. In this work, we present an approach to\r\ndetecting COVID-19 in chest X-rays that combines reconstruction-based anomaly discov-\r\nery with perturbation-based attribution. Specifically, we use a Variational Autoencoder\r\ntrained on healthy lungs to identify lung anomalies, and we additionally evaluate a Per-\r\nceptual Autoencoder (PAE) that incorporates a perceptual loss to produce sharper recon-\r\nstructions and more contrastive residual maps. These signals are then used to highlight\r\ncritical image evidence through both patch-level scoring (grid-based occlusion) and pixel-\r\nlevel masking derived from reconstruction-error maps, enabling healthcare professionals\r\nto localize relevant regions more effectively. Moreover, the proposed framework provides\r\nclearer explanations of the model\u2019s decisions by quantifying the prediction change after\r\noccluding the identified regions, reducing the complexity of the \u201cblack boxes\u201d generated\r\nby deep learning neural networks. With this methodology, we aim to improve the effec-\r\ntiveness and reliability of early COVID-19 detection through chest X-rays.<\/jats:p>","DOI":"10.19153\/cleiej.29.3.6","type":"journal-article","created":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T18:41:24Z","timestamp":1784054484000},"source":"Crossref","is-referenced-by-count":0,"title":["Explainable COVID-19 Classification Via Variational and Perceptual Autoencoder-Guided Occlusion"],"prefix":"10.19153","volume":"29","author":[{"given":"Rodrigo","family":"Bayuk","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Joel","family":"Manquel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Orietta","family":"Nicolis","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Billy","family":"Peralta","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Luis","family":"Caro","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"8231","published-online":{"date-parts":[[2026,7,14]]},"container-title":["CLEI Electronic Journal"],"original-title":[],"link":[{"URL":"https:\/\/clei.org\/cleiej\/index.php\/cleiej\/article\/download\/1032\/584","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/clei.org\/cleiej\/index.php\/cleiej\/article\/download\/1032\/584","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T18:41:25Z","timestamp":1784054485000},"score":1,"resource":{"primary":{"URL":"https:\/\/clei.org\/cleiej\/index.php\/cleiej\/article\/view\/1032"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,14]]},"references-count":0,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2026,7,14]]}},"URL":"https:\/\/doi.org\/10.19153\/cleiej.29.3.6","relation":{},"ISSN":["0717-5000"],"issn-type":[{"value":"0717-5000","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,14]]}}}