Journal: Frontiers in Medicine

Article Title: Application of deep learning convolutional neural networks to identify gastric squamous cell carcinoma in mice

doi: 10.3389/fmed.2025.1587417

Figure Lengend Snippet: The background and tissue regions of slides were identified based on the HALO AI Tissue Detection-BF algorithm model. (A) The original pathological data image; (B) The result of the recognition based on Tissue Detection BF algorithm model, with the tissue area being green and the slide background being gray.

Article Snippet: To establish an automated tumor recognition deep learning model, two pathologists manually annotated the tumor regions (primarily refer to areas dominated by squamous epithelial tumor cells) and non-tumor regions (primarily refer to blank areas, stromal regions and necrosis areas) in selected samples of forestomach using the HALO AI Tissue Pathology Image Analysis System (V3.6.4134, Indica Labs, Inc., Albuquerque, NM).

Techniques:

Journal: Frontiers in Medicine

Article Title: Application of deep learning convolutional neural networks to identify gastric squamous cell carcinoma in mice

doi: 10.3389/fmed.2025.1587417

Figure Lengend Snippet: The artifact regions were identified in tissues based on the HALO AI QC slide algorithm model. (A) The original pathological data images and dusts. (B) High-magnification micrograph of dusts. (C,D) The artifacts identified by HALO AI, with red areas indicating dust artifact.

Article Snippet: To establish an automated tumor recognition deep learning model, two pathologists manually annotated the tumor regions (primarily refer to areas dominated by squamous epithelial tumor cells) and non-tumor regions (primarily refer to blank areas, stromal regions and necrosis areas) in selected samples of forestomach using the HALO AI Tissue Pathology Image Analysis System (V3.6.4134, Indica Labs, Inc., Albuquerque, NM).

Techniques:

Journal: Frontiers in Medicine

Article Title: Application of deep learning convolutional neural networks to identify gastric squamous cell carcinoma in mice

doi: 10.3389/fmed.2025.1587417

Figure Lengend Snippet: Automatic recognition of the normal epithelial tissue and other tissue based on DenseNet AI of HALO AI. (A–C) The original pathological images of three normal mouse gastric tissues. (D–F) Correspond to HALO automatically annotating images of three normal gastric tissues (A–C) , respectively. Red represents the normal epithelial tissue automatically recognized by HALO, while green represents normal other tissues.

Article Snippet: To establish an automated tumor recognition deep learning model, two pathologists manually annotated the tumor regions (primarily refer to areas dominated by squamous epithelial tumor cells) and non-tumor regions (primarily refer to blank areas, stromal regions and necrosis areas) in selected samples of forestomach using the HALO AI Tissue Pathology Image Analysis System (V3.6.4134, Indica Labs, Inc., Albuquerque, NM).

Techniques:

Journal: Frontiers in Medicine

Article Title: Application of deep learning convolutional neural networks to identify gastric squamous cell carcinoma in mice

doi: 10.3389/fmed.2025.1587417

Figure Lengend Snippet: Tumors area manually annotated by pathologist. (A) The original pathological data image. (B) Green is tumor area and red is the tumor area automatically annotated by HALO AI. (C) Black is the tumor area manually revised and annotated by pathologist.

Article Snippet: To establish an automated tumor recognition deep learning model, two pathologists manually annotated the tumor regions (primarily refer to areas dominated by squamous epithelial tumor cells) and non-tumor regions (primarily refer to blank areas, stromal regions and necrosis areas) in selected samples of forestomach using the HALO AI Tissue Pathology Image Analysis System (V3.6.4134, Indica Labs, Inc., Albuquerque, NM).

Techniques: