Visual Pattern Recognition Networks (VPRN) delivers high performance using deep neural networks. With the advancements in deep neural networks VPR network has gained wide popularity. Continuous advancements will be nurtured with the availability of big data and enormous computing powers. However, such DNN based VPRN models are plunged with computational complexities, intense memory requirements, huge energy expenses which impedes its deployment in resource constrained, strict latency required environments such as edgeAI. For instance, the VGG-16 model needs 500 MB of storage space, has 138 million parameters and involves 15.5 billion Floating Point Operations (FLOPs) to classify a single image with a 32 bit floating point addition that consumes 0.9 pJ. Such overheads demand for compression of VPRN models without impairing its performance. Various compression methods are reported in literature. This survey paper presents a survey on pruning, a popular compression technique for DNN as applied for VPRN. This paper presents comprehensive survey, comparison and points further research direction. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.