This is A Flask App that uses DRÆM (Discriminatively Trained Reconstruction Embedding for Surface Anomaly Detection) for detecting anomalies in surfaces, such as defects or damage, using a combination of reconstruction and classification techniques. For now the app caters only one class that is 'Fabric', but I'll be adding all the classes present in MVtec Dataset.
Please download the model files and place inside static/models/
Model files can be downloaded from here
