Considerations Of Integrating Computing-In-Memory And Processing-In-Sensor Into Convolutional Neural Network Accelerators For Low-Power Edge Devices | IEEE Conference Publication | IEEE Xplore