M
Muhammad Mohsin Ghaffar
Researcher at Kaiserslautern University of Technology
Publications - 10
Citations - 114
Muhammad Mohsin Ghaffar is an academic researcher from Kaiserslautern University of Technology. The author has contributed to research in topics: Historical document & Optical character recognition. The author has an hindex of 3, co-authored 10 publications receiving 62 citations.
Papers
More filters
Proceedings ArticleDOI
FINN-L: Library Extensions and Design Trade-Off Analysis for Variable Precision LSTM Networks on FPGAs
Vladimir Rybalkin,Alessandro Pappalardo,Muhammad Mohsin Ghaffar,Giulio Gambardella,Norbert Wehn,Michaela Blott +5 more
TL;DR: This paper presents the first systematic exploration of this design space as a function of precision for Bidirectional Long Short-Term Memory (BiLSTM) neural network, and provides the first open source HLS library extension of FINN for parameterizable hardware architectures of LSTM layers on FPGAs which offers full precision flexibility and allows for parameterized performance scaling.
Proceedings ArticleDOI
An In-DRAM Neural Network Processing Engine
Chirag Sudarshan,Jan Lappas,Muhammad Mohsin Ghaffar,Vladimir Rybalkin,Christian Weis,Matthias Jung,Norbert Wehn +6 more
TL;DR: This work presents a novel PIM based Binary Weighted Network (BWN) inference accelerator design that is inline with the commodity Dynamic Random Access Memory (DRAM) design and process and is extremely energy efficient.
Journal ArticleDOI
AI-ForestWatch: semantic segmentation based end-to-end framework for forest estimation and change detection using multi-spectral remote sensing imagery
Annus Zulfiqar,Muhammad Mohsin Ghaffar,Muhammad Shahzad,Christian Weis,Muhammad Imran Malik,Faisal Shafait,Norbert Wehn +6 more
TL;DR: In this paper, the authors presented an end-to-end framework for forest estimation and change analysis using deep convolution neural network-based semantic segmentation to process multi-spectral space-borne images to quantitatively monitor the forest cover change patterns.
Proceedings ArticleDOI
A Low Power In-DRAM Architecture for Quantized CNNs using Fast Winograd Convolutions
TL;DR: In this article, the authors proposed a novel DRAM-based PIM architecture for quantized (8-bit weight and input) CNN inference by utilizing the complexity reduction offered by fast convolution algorithms.
Proceedings ArticleDOI
iDocChip: A Configurable Hardware Architecture for Historical Document Image Processing: Percentile Based Binarization
Vladimir Rybalkin,Syed Saqib Bukhari,Muhammad Mohsin Ghaffar,Aqib Ghafoor,Norbert Wehn,Andreas Dengel +5 more
TL;DR: This paper presents a hybrid hardware-software FPGA-based accelerator that outperforms the existing anyOCR software implementation running on i7-4790T in terms of runtime by factor of 21, while achieving energy efficiency of 10 Images/J that is higher than that achieved by low power embedded processors with negligible loss of recognition accuracy.