{"article":{"annotation":{"article_id":"mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026","available_at":"2026-10-20T14:20:50.997937Z","geography_revision":"ef6ee0010dda5fe7b240fd514749b43e214c57da73222d2e40ec83b276dd0977","id":"70b1cddaf5b21d5be8669932db4fda8fc67f568ebb402b806448f7337c2d72ec","input_sha256":"618b1e4125a8eec188bc34454886c86ead7ace94c0c8f476f275eb57042ca34b","limitations":["Rule-derived subjects are candidates, not verified editorial labels or calibrated model probabilities.","Coverage is distinct from event location. Unknown country/city/coordinates remain absent; no automatic geocoding.","Article geometries keep their stated precision and methodology; reproducing them is not independent coordinate verification."],"method":"controlled-rules-v1","places":[{"admin1":null,"city":null,"coordinate_note":null,"coordinate_reference_system":"OGC:CRS84","coordinate_source":null,"country":"AE","evidence":[{"field":"country","language":"en","note":"Explicit preserved editorial country/region; no city or national centroid inferred","source_urls":["https://mbzuai.ac.ae/news-events/news/3d-medical-ai-can-read-ct-scans-new-benchmark-asks-whether-it-knows-where-things"]}],"geometry":null,"id":"iso3166-1:AE","kind":"country-or-area","labels":{"ar":"الإمارات العربية المتحدة","en":"United Arab Emirates","es":"Emiratos Árabes Unidos","fr":"Émirats arabes unis"},"match_status":"candidate","precision":"regional","role":"coverage","uncertainty_m":null}],"rules_sha256":"fb5ad3c27c91cb840f7a0a05f7f64a5cb26ad0374409ff9457f277c7092e22b4","schema_version":1,"subjects":[{"concept_id":"medtop:20001298","confidence":null,"confidence_method":null,"evidence":[{"field":"en.title+en.summary","language":"en","note":"Subject suggested from the title or summary: AI","source_urls":["https://mbzuai.ac.ae/news-events/news/3d-medical-ai-can-read-ct-scans-new-benchmark-asks-whether-it-knows-where-things"]}],"method":"controlled-rules-v1","rule_id":"artificial-intelligence","status":"candidate"},{"concept_id":"medtop:07000000","confidence":null,"confidence_method":null,"evidence":[{"field":"en.title+en.summary","language":"en","note":"Subject suggested from the title or summary: medical","source_urls":["https://mbzuai.ac.ae/news-events/news/3d-medical-ai-can-read-ct-scans-new-benchmark-asks-whether-it-knows-where-things"]}],"method":"controlled-rules-v1","rule_id":"health","status":"candidate"}],"taxonomy_revision":"1f2db31d811fca07e99b16d402c10e2cc2965c9df1c7b41d9be348b40a24e7fb","taxonomy_version":"IPTC Media Topics 2026-Q2"},"api_url":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026?lang=ar","category":"technology","content":{"key_points":[],"paragraphs":["طور باحثون في جامعة MBZUAI معيار CT-SpatialVQA لاختبار قدرة أنظمة الذكاء الاصطناعي الطبي ثلاثية الأبعاد على تتبع العلاقات المكانية للتشريح والاختلالات عبر أحجام التصوير المقطعي المحوسب. وذكر تقرير للجامعة في 19 أكتوبر أن مجموعة البيانات تضم 9,077 زوجاً من الأسئلة والأجوبة المكانية مشتقة من 1,601 فحصاً مقطعياً للصدر وتقارير الأشعة.","قيّم الباحثون ثمانية نماذج للرؤية واللغة في إعداد دون أمثلة تدريبية خاصة بالمهمة. ولم يحقق أي منها دقة 50%، فيما بلغ متوسط النماذج الثمانية 34.08%. وسجل CT-Chat نسبة 43.69%، وMedEvalKit نسبة 40.16%، وMedGemma-1.5 نسبة 38.24%، بينما سجل VILA-M3 نسبة 28.20%. وقُبل العمل كعرض شفهي في MICCAI 2026. وهذه نتائج معيار بحثي محدد ولا تثبت الكفاءة السريرية أو الاستخدام المعتمد في رعاية المرضى."],"sections":[]},"corrections":[],"country":"AE","data":{"blocks":[{"columns":[{"key":"model","label":{"ar":"النموذج","de":"Modell","en":"Model","es":"Modelo","fr":"Modèle","pt":"Modelo","tr":"Model"},"language":"en","type":"text"},{"key":"value","label":{"ar":"الدقة","de":"Genauigkeit","en":"Accuracy","es":"Precisión","fr":"Précision","pt":"Precisão","tr":"Doğruluk"},"type":"number","unit":"percent"},{"key":"unit","label":{"ar":"الوحدة","de":"Einheit","en":"Unit","es":"Unidad","fr":"Unité","pt":"Unidade","tr":"Birim"},"language":"en","type":"text"}],"description":{"ar":"نتائج مختارة ومتوسط عام من تقييم الجامعة لثمانية نماذج.","de":"Ausgewählte Ergebnisse und Gesamtdurchschnitt aus der Prüfung von acht Modellen durch die Universität.","en":"Selected results and overall mean from the university’s evaluation of eight models.","es":"Resultados seleccionados y promedio general de la evaluación universitaria de ocho modelos.","fr":"Résultats sélectionnés et moyenne générale de l’évaluation de huit modèles par l’université.","pt":"Resultados selecionados e média geral da avaliação universitária de oito modelos.","tr":"Üniversitenin bildirdiği sekiz model testinden örnek sonuçlar ve genel ortalama."},"id":"ct-spatialvqa-model-accuracies","methodology":{"ar":"هذه نتائج دقة دون أمثلة تدريبية خاصة على معيار CT-SpatialVQA. ويشمل المتوسط ثمانية نماذج؛ ويعرض الجدول أربعة نماذج ذكرها المصدر منفردة والمتوسط. ولا تمثل النتائج تحققاً سريرياً.","de":"Dies sind Zero-Shot-Genauigkeitswerte des CT-SpatialVQA-Benchmarks. Der Durchschnitt umfasst acht Modelle; die Tabelle zeigt vier einzeln genannte Modelle und den Durchschnitt. Es handelt sich nicht um eine klinische Validierung.","en":"These are zero-shot accuracy results on the CT-SpatialVQA benchmark. The mean covers eight models; the table lists four models reported individually in the source and the mean. Results are not clinical validation.","es":"Son resultados de precisión sin ejemplos en el benchmark CT-SpatialVQA. El promedio abarca ocho modelos; la tabla presenta cuatro modelos detallados por la fuente y el promedio. No es una validación clínica.","fr":"Il s’agit de résultats de précision zero-shot sur le benchmark CT-SpatialVQA. La moyenne porte sur huit modèles ; le tableau présente les quatre modèles détaillés dans la source et la moyenne. Ces résultats ne valent pas validation clinique.","pt":"São resultados de precisão zero-shot no benchmark CT-SpatialVQA. A média inclui oito modelos; a tabela apresenta quatro modelos discriminados pela fonte e a média. Não se trata de validação clínica.","tr":"Bunlar CT-SpatialVQA kıyaslamasındaki sıfır örnekli doğruluk sonuçlarıdır. Ortalamaya sekiz model dahildir; tabloda yalnızca kaynakta ayrı ayrı belirtilen dört model ile ortalama gösterilmiştir. Sonuçlar klinik doğrulama değildir."},"rowTranslations":[{"unit":{"ar":"بالمئة","de":"Prozent","en":"percent","es":"por ciento","fr":"pour cent","pt":"por cento","tr":"yüzde"}},{"unit":{"ar":"بالمئة","de":"Prozent","en":"percent","es":"por ciento","fr":"pour cent","pt":"por cento","tr":"yüzde"}},{"unit":{"ar":"بالمئة","de":"Prozent","en":"percent","es":"por ciento","fr":"pour cent","pt":"por cento","tr":"yüzde"}},{"unit":{"ar":"بالمئة","de":"Prozent","en":"percent","es":"por ciento","fr":"pour cent","pt":"por cento","tr":"yüzde"}},{"model":{"ar":"متوسط ثمانية نماذج","de":"Durchschnitt der acht Modelle","en":"Average across eight models","es":"Promedio de los ocho modelos","fr":"Moyenne des huit modèles","pt":"Média dos oito modelos","tr":"Sekiz modelin ortalaması"},"unit":{"ar":"بالمئة","de":"Prozent","en":"percent","es":"por ciento","fr":"pour cent","pt":"por cento","tr":"yüzde"}}],"rows":[{"model":"CT-Chat","unit":"percent","value":43.69},{"model":"MedEvalKit","unit":"percent","value":40.16},{"model":"MedGemma-1.5","unit":"percent","value":38.24},{"model":"VILA-M3","unit":"percent","value":28.2},{"model":"Average across eight models","unit":"percent","value":34.08}],"sourceUrls":["https://mbzuai.ac.ae/news-events/news/3d-medical-ai-can-read-ct-scans-new-benchmark-asks-whether-it-knows-where-things"],"title":{"ar":"نتائج دقة النماذج في CT-SpatialVQA","de":"Genauigkeit der Modelle bei CT-SpatialVQA","en":"CT-SpatialVQA model accuracy results","es":"Resultados de precisión de modelos en CT-SpatialVQA","fr":"Résultats de précision des modèles CT-SpatialVQA","pt":"Resultados de precisão dos modelos no CT-SpatialVQA","tr":"CT-SpatialVQA model doğruluk sonuçları"},"type":"table","updatedAt":"2026-10-20T14:12:00Z"},{"columns":[{"key":"item","label":{"ar":"العنصر","de":"Bestandteil","en":"Component","es":"Componente","fr":"Élément","pt":"Componente","tr":"Bileşen"},"language":"en","type":"text"},{"key":"value","label":{"ar":"العدد","de":"Anzahl","en":"Count","es":"Cantidad","fr":"Nombre","pt":"Número","tr":"Sayı"},"type":"number","unit":"count"},{"key":"unit","label":{"ar":"الوحدة","de":"Einheit","en":"Unit","es":"Unidad","fr":"Unité","pt":"Unidade","tr":"Birim"},"language":"en","type":"text"}],"description":{"ar":"نطاق الأسئلة والفحوصات في معيار البحث.","de":"Umfang der Fragen und Aufnahmen im Forschungsbenchmark.","en":"Question and scan coverage in the research benchmark.","es":"Cobertura de preguntas y escáneres en la prueba.","fr":"Nombre de questions et d’examens dans le benchmark.","pt":"Cobertura de perguntas e exames no benchmark.","tr":"Araştırma kıyaslamasının soru ve görüntü kapsamı."},"id":"ct-spatialvqa-dataset-size","methodology":{"ar":"تصف الأعداد معيار الاختبار البحثي، لا مجموعة مرضى أو أداء تشخيصياً أو نتائج سريرية. ويذكر المصدر أن الأسئلة البالغ عددها 9,077 اشتُقت من 1,601 فحصاً مقطعياً للصدر وتقارير أشعة في مجموعة الاختبار.","de":"Die Zahlen beschreiben den Forschungsbenchmark, keine Patientenpopulation, Diagnoseleistung oder klinischen Behandlungsergebnisse. Laut Quelle wurden die 9.077 Fragen aus 1.601 Thorax-CT-Aufnahmen und Befunden im Testsatz abgeleitet.","en":"Counts describe the researchers’ test benchmark, not a patient population, diagnostic performance or clinical outcome. The source says the 9,077 questions were derived from 1,601 chest CT scans and radiology reports in the test set.","es":"Las cifras describen la prueba de investigación, no una población de pacientes, un rendimiento diagnóstico ni resultados clínicos. La fuente indica que las 9.077 preguntas proceden de 1.601 tomografías torácicas e informes radiológicos.","fr":"Ces nombres décrivent le benchmark de test des chercheurs, pas une population de patients, une performance diagnostique ni des résultats cliniques. La source indique que les 9 077 questions proviennent de 1 601 scanners thoraciques et de leurs rapports.","pt":"Os valores descrevem o benchmark de investigação, não uma população de doentes, desempenho diagnóstico ou resultados clínicos. A fonte afirma que as 9.077 perguntas derivam de 1.601 exames torácicos e relatórios radiológicos do conjunto de teste.","tr":"Sayılar araştırmacıların kurduğu test kıyaslamasını tanımlar; hasta nüfusu, tanı performansı veya klinik sonuç ölçümü değildir. Kaynak 9.077 sorunun test kümesindeki 1.601 göğüs BT taramasından ve radyoloji raporlarından türetildiğini belirtir."},"rowTranslations":[{"item":{"ar":"أزواج أسئلة وأجوبة مكانية","de":"Räumliche Frage-Antwort-Paare","en":"Spatial question-answer pairs","es":"Pares de preguntas y respuestas espaciales","fr":"Paires de questions-réponses spatiales","pt":"Pares de perguntas e respostas espaciais","tr":"Mekânsal soru-cevap çiftleri"},"unit":{"ar":"عدد","de":"Anzahl","en":"count","es":"cantidad","fr":"nombre","pt":"quantidade","tr":"adet"}},{"item":{"ar":"فحوص التصوير المقطعي للصدر","de":"Thorax-CT-Aufnahmen","en":"Chest CT scans","es":"Tomografías computarizadas de tórax","fr":"Examens de tomodensitométrie thoracique","pt":"Exames de tomografia computadorizada do tórax","tr":"Göğüs BT taramaları"},"unit":{"ar":"عدد","de":"Anzahl","en":"count","es":"cantidad","fr":"nombre","pt":"quantidade","tr":"adet"}}],"rows":[{"item":"Spatial question-answer pairs","unit":"count","value":9077},{"item":"Chest CT scans","unit":"count","value":1601}],"sourceUrls":["https://mbzuai.ac.ae/news-events/news/3d-medical-ai-can-read-ct-scans-new-benchmark-asks-whether-it-knows-where-things"],"title":{"ar":"مجموعة بيانات CT-SpatialVQA","de":"CT-SpatialVQA-Datensatz","en":"CT-SpatialVQA dataset","es":"Conjunto de datos CT-SpatialVQA","fr":"Jeu de données CT-SpatialVQA","pt":"Conjunto de dados CT-SpatialVQA","tr":"CT-SpatialVQA veri kümesi"},"type":"table","updatedAt":"2026-10-20T14:12:00Z"}],"version":1},"data_downloads":[{"formats":{"csv":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026/data/ct-spatialvqa-model-accuracies?lang=ar&format=csv","json":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026/data/ct-spatialvqa-model-accuracies?lang=ar&format=json"},"id":"ct-spatialvqa-model-accuracies"},{"formats":{"csv":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026/data/ct-spatialvqa-dataset-size?lang=ar&format=csv","json":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026/data/ct-spatialvqa-dataset-size?lang=ar&format=json"},"id":"ct-spatialvqa-dataset-size"}],"data_types":["table"],"editorial_illustration":{"alt":"رسم توضيحي مفاهيمي للتكنولوجيا والمعالجات والمعلومات المترابطة.","asset_id":"newsverum-v1-technology","avif_variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.avif","width":480},{"height":480,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-720.avif","width":720},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.avif","width":960},{"height":800,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1200.avif","width":1200},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.avif","width":1536}],"credit":"Newsverum","documentary":false,"height":1024,"kind":"symbolic-editorial-illustration","language":"ar","scope":"topic","url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.webp","width":480},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.webp","width":960},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","width":1536}],"width":1536},"event_id":"event-446cd5a4-c43f-50a6-9009-faa697af4485","event_period":"2026-10-19","event_period_basis":"event-or-reference-period","id":"mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026","image":null,"importance":0.6,"language":"ar","modified_at":"2026-10-20T14:20:10+00:00","publication_basis":"publication-record","published_at":"2026-10-20T14:20:10+00:00","reading_minutes":1,"region":"world","relations":[],"section_illustration":{"alt":"رسم توضيحي مفاهيمي للتكنولوجيا والمعالجات والمعلومات المترابطة.","asset_id":"newsverum-v1-technology","avif_variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.avif","width":480},{"height":480,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-720.avif","width":720},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.avif","width":960},{"height":800,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1200.avif","width":1200},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.avif","width":1536}],"credit":"Newsverum","documentary":false,"height":1024,"kind":"symbolic-editorial-illustration","language":"ar","scope":"section","url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.webp","width":480},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.webp","width":960},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","width":1536}],"width":1536},"slug":"mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026","sources":[{"checkedAt":"2026-10-20","name":"Mohamed bin Zayed University of Artificial Intelligence","publishedAt":"2026-10-19","url":"https://mbzuai.ac.ae/news-events/news/3d-medical-ai-can-read-ct-scans-new-benchmark-asks-whether-it-knows-where-things"}],"summary":"يختبر معيار CT-SpatialVQA من MBZUAI قدرة نماذج الذكاء الاصطناعي الطبي على تحديد العلاقات المكانية في فحوصات CT.","text":"يختبر معيار CT-SpatialVQA من MBZUAI قدرة نماذج الذكاء الاصطناعي الطبي على تحديد العلاقات المكانية في فحوصات CT.\n\nطور باحثون في جامعة MBZUAI معيار CT-SpatialVQA لاختبار قدرة أنظمة الذكاء الاصطناعي الطبي ثلاثية الأبعاد على تتبع العلاقات المكانية للتشريح والاختلالات عبر أحجام التصوير المقطعي المحوسب. وذكر تقرير للجامعة في 19 أكتوبر أن مجموعة البيانات تضم 9,077 زوجاً من الأسئلة والأجوبة المكانية مشتقة من 1,601 فحصاً مقطعياً للصدر وتقارير الأشعة.\n\nقيّم الباحثون ثمانية نماذج للرؤية واللغة في إعداد دون أمثلة تدريبية خاصة بالمهمة. ولم يحقق أي منها دقة 50%، فيما بلغ متوسط النماذج الثمانية 34.08%. وسجل CT-Chat نسبة 43.69%، وMedEvalKit نسبة 40.16%، وMedGemma-1.5 نسبة 38.24%، بينما سجل VILA-M3 نسبة 28.20%. وقُبل العمل كعرض شفهي في MICCAI 2026. وهذه نتائج معيار بحثي محدد ولا تثبت الكفاءة السريرية أو الاستخدام المعتمد في رعاية المرضى.","title":"معيار من MBZUAI يختبر الاستدلال المكاني في الذكاء الاصطناعي الطبي ثلاثي الأبعاد","topics":[],"translations":[{"api_url":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026?lang=tr","language":"tr","url":"https://newsverum.com/tr/news/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026"},{"api_url":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026?lang=en","language":"en","url":"https://newsverum.com/en/news/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026"},{"api_url":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026?lang=ar","language":"ar","url":"https://newsverum.com/ar/news/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026"},{"api_url":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026?lang=fr","language":"fr","url":"https://newsverum.com/fr/news/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026"},{"api_url":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026?lang=es","language":"es","url":"https://newsverum.com/es/news/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026"},{"api_url":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026?lang=de","language":"de","url":"https://newsverum.com/de/news/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026"},{"api_url":"https://newsverum.com/api/v1/articles/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026?lang=pt","language":"pt","url":"https://newsverum.com/pt/news/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026"}],"url":"https://newsverum.com/ar/news/mbzuai-ct-spatialvqa-3d-medical-ai-benchmark-2026"},"version":"1"}