Image classification using limited labeled data

A non-transitory program storage device readable by a computer, tangibly embodying a program of instructions executed by the computer to perform the method steps for digital image classification and localization, the method comprising the steps of:

• receiving a digital image of a biological organism from an imaging apparatus, said digital image comprising a plurality of intensities on a 2-dimensional (2D) grid of points;
• generating a plurality of discriminative representations of the received digital image by extracting dominant characteristics of the image from three different viewpoints, wherein the plurality of discriminative representations form a 3-dimensional (3D) digital image;
• combining the 3D digital image with the received 2D digital image in a convolutional neural network that outputs a 3-channel feature map that localizes image abnormalities in each of the three channels and includes a detection confidence that each abnormalities is a neoplasm;
• providing the 3-channel feature map to a controller of a robotic surgical device wherein the robotic surgical device uses the 3-channel feature map to locate the neoplasm within the biological organism in a surgical procedure for treating the neoplasm.