A sparse representation of the pathologist's interaction with Whole Slide Images in order to improve the captured relevance in regions of interest
Cargando...
Autores
Santiago Garnica, Daniel Ernesto
Director
Tipo de contenido
Document language:
Español
Fecha
Título de la revista
ISSN de la revista
Título del volumen
Documentos PDF
Resumen
Durante una tarea de diagn´ostico, un pat´ologo corre sobre un WSI (”Whole Slide Image”), con el objetivo de descubrir patrones patol´ogicos relevantes. Un microscopio virtual capta estas estructuras pero tambi´en otros patrones celulares con diferentes o ning´un significado diagn´ostico. La anotaci´on de estas im´agenes depende entonces de la delineaci´on manual, una tarea dif´ıcil en la pr´actica. Este art´ıculo presenta un nuevo m´etodo para detectar regiones relevantes en WSI utilizando las navegaciones de rutina en un microscopio virtual. El m´etodo construye una representaci´on escasa o un diccionario de cada trayecto de navegaci´on y determina la relevancia oculta maximizando la incoherencia entre varios trayectos. Los diccionarios obtenidos se proyectan entre s´ı y la informaci´on relevante se establece en los ´atomos del diccionario cuya similitud es superior a un umbral personalizado. La evaluaci´on se realiz´o con 6 imagenes patologicas con n´ucleos segmentados a partir de una biopsia cut´anea diagnosticada con carcinoma basocelular (BCC). Los resultados muestran que nuestro enfoque supera a la l´ınea base en m´as de 20 %
Abstract: During a diagnosis task, a Pathologist looks over a WSI (”Whole Slide Image”), aiming to find out relevant pathological patterns. Nonetheless, a virtual microscope captures these structures, but also other cellular patterns with different or none diagnostic meaning. Annotation of these images depends on manual delineation, which in practice becomes a hard task. This article contributes a new method for detecting relevant regions in WSI using the routine navigations in a virtual microscope. This method constructs a sparse representation or dictionary of each navigation path and determines the hidden relevance by maximizing the incoherence between several paths. The resulting dictionaries are then projected onto each other and relevant information is set to the dictionary atoms whose similarity is higher than a custom threshold. Evaluation was performed with 6 pathological images segmented from a skin biopsy already diagnosed with basal cell carcinoma (BCC). Results show that our proposal outperforms the baseline by more than 20 %.
Abstract: During a diagnosis task, a Pathologist looks over a WSI (”Whole Slide Image”), aiming to find out relevant pathological patterns. Nonetheless, a virtual microscope captures these structures, but also other cellular patterns with different or none diagnostic meaning. Annotation of these images depends on manual delineation, which in practice becomes a hard task. This article contributes a new method for detecting relevant regions in WSI using the routine navigations in a virtual microscope. This method constructs a sparse representation or dictionary of each navigation path and determines the hidden relevance by maximizing the incoherence between several paths. The resulting dictionaries are then projected onto each other and relevant information is set to the dictionary atoms whose similarity is higher than a custom threshold. Evaluation was performed with 6 pathological images segmented from a skin biopsy already diagnosed with basal cell carcinoma (BCC). Results show that our proposal outperforms the baseline by more than 20 %.