XCAT - Lightweight Quantized Single Image Super-Resolution Using Heterogeneous Group Convolutions and Cross Concatenation

Mustafa Ayazoglu, Bahri Batuhan Bilecen

Abstract: We propose a lightweight, single-image super-resolution mobile device network named XCAT, and introduce Heterogeneous Group Convolution Blocks with Cross Concatenations (HXBlock). The heterogeneous split of the input channels to the group convolution blocks reduces the number of operations, and cross concatenation allows for information flow between the intermediate input tensors of cascaded HXBlocks. Cross concatenations inside HXBlocks can also avoid using more expensive operations like 1  $$\times $$  1 convolutions. To further prevent expensive tensor copy operations, XCAT utilizes non-trainable convolution kernels to apply upsampling operations. Designed with integer quantization in mind, XCAT also utilizes several techniques in training, like intensity-based data augmentation. Integer quantized XCAT operates in real-time on Mali-G71 MP2 GPU with 320 ms, and on Synaptics Dolphin NPU with 30 ms (NCHW) and 8.8 ms (NHWC), suitable for real-time applications.