Please refer to the article related to this dataset: Deep learning enables the automation of grading histological tissue engineered cartilage images for quality control standardization https://doi.org/10.1016/j.joca.2020.12.018 The objective was to automate and standardize the grading of histological images of in vitro engineered cartilage tissues using deep learning. Cartilaginous tissues were engineered from human articular or nasal chondrocytes or from bone-marrow derived mesenchymal stromal cells. Safranin O and fast green stained histological images of the tissues were graded for chondrogenic quality according to a modified version of the Bern score (Modified Bern Score), which ranks images on a scale from 0 to 6 according to the intensity of staining and cell morphology. The images were graded by two experts and grouped into four categories with the following grades: 0, 1-2, 3-4, and 5-6. Deep learning with transfer learning was used to train a model to classify images into these histological score groups. Cartilaginous tissues of diverse quality were produced, covering all categories of the Modified Bern Score. Transfer learning using a pretrained DenseNet model was selected.
Steps to reproduce
The "AutomatedGrading.ipynb" JupyterNotebook file finds the labeled images in the pictures directory and scores all the images within. The "MBS_For_Mac.ipynb" JupyterNotebook file visualizes the scores given to one image by the model, testpic.tif is provided as an example here. Both notebooks use the model weights stored in the state_dict_densenet_classifier1.pt file.