We introduce a new visualisation technique for CNNs called Principal Feature Visualisation (PFV). It uses a single forward pass of the original network to map principal features from the nal convolutional layer to the original image space as RGB channels. By working on a batch of images we can...

We consider the case of robotic bin picking of reflective steel parts, using a structured light 3D camera as a depth imaging device. In this paper, we present a new method for bin picking, based on a dual-resolution convolutional neural network trained entirely in a simulated environment. The dual...

The objectives of this project have been to image bones in whole fish and fillets in 9 different spices and to provide detailed information about the size, orientation and location of pinbones and the walking stick bone in fillets. For each spices 2-4 fillets were CT scanned and analyzed. The bones...