ZipNet: ZFNet-level Accuracy with 48× Fewer Parameters

Arren Matthew C. Antioquia, Daniel Stanley Tan, Arnulfo Azcarraga, Wen-Huang Cheng, Kai-Lung Hua

Research output: Chapter in Book/Report/Conference proceedingConference Article in proceedingAcademicpeer-review


With the introduction of Convolutional Neural Networks, models for image classification achieve higher classification accuracy. Based on the pattern of the design of CNN architectures, increasing the number of layers equates to a higher classification accuracy, but also increases the number of parameters and model size. This negatively affects the model training time, processing time, and memory requirement. We develop ZipNet, a CNN architecture with a higher classification accuracy than ZFNet, the winner of ILSVRC 2013, but with 48.5× smaller model size and 48.7× fewer parameters. The classification accuracy of ZipNet is higher than the performance of ZFNet and SqueezeNet on all configurations of the Caltech-256 dataset with varying number of training examples.
Original languageEnglish
Title of host publication2018 IEEE Visual Communications and Image Processing (VCIP)
Subtitle of host publicationProceedings
ISBN (Electronic)978-1-5386-4458-4
ISBN (Print)978-1-5386-4459-1
Publication statusPublished - 2 Jul 2018
Externally publishedYes
EventIEEE Visual Communications and Image Processing - Taichung, Taiwan, Province of China
Duration: 9 Dec 201812 Dec 2018


ConferenceIEEE Visual Communications and Image Processing
Abbreviated titleVCIP 2018
Country/TerritoryTaiwan, Province of China
Internet address


  • Convolutional Neural Networks
  • Deep Learning
  • Image Classification
  • Model Compression
  • Object Classification


Dive into the research topics of 'ZipNet: ZFNet-level Accuracy with 48× Fewer Parameters'. Together they form a unique fingerprint.

Cite this