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

TL;DR

XCAT is a lightweight, quantized super-resolution network optimized for mobile devices, using heterogeneous group convolutions and cross concatenation to reduce computation and enable real-time performance.

Contribution

The paper introduces HXBlock with heterogeneous group convolutions and cross concatenation, improving efficiency and performance for mobile super-resolution.

Findings

Operates in real time on Mali-G71 GPU (320ms)

Achieves 30ms on Synaptics Dolphin NPU (NCHW)

Suitable for real-time mobile applications

Abstract

We propose a lightweight, single image super-resolution network for mobile devices, named XCAT. XCAT introduces 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 1x1 convolutions. To further prev ent expensive tensor copy operations, XCAT utilizes non-trainable convolution kernels to apply up sampling operations. Designed with integer quantization in mind, XCAT also utilizes several techniques on training, like intensity-based data augmentation. Integer quantized XCAT operates in real time on Mali-G71 MP2 GPU with 320ms, and on Synaptics Dolphin NPU with 30ms (NCHW) and 8.8ms (NHWC), suitable for real-time applications.