Energy Efficient Approximate 3D Image Reconstruction
Published in IEEE Transactions on Emerging Topics in Computing, 2022
This paper demonstrates an efficient and accelerated parallel, sparse depth reconstruction framework using compressed sensing (CS) and approximate computing (AC) techniques. The proposed framework aims to achieve energy efficiency in 3D image reconstruction by exploiting the inherent sparsity in depth data and applying approximate computing methods to reduce computational complexity.
Key contributions include:
Compressed Sensing: Utilization of CS techniques to reconstruct 3D images from fewer measurements, leveraging the sparsity of the data.
Approximate Computing: Application of AC methods to approximate certain computations, reducing energy consumption without significantly compromising the quality of the reconstructed images.
Parallel Processing: Implementation of the framework on parallel processing architectures to accelerate the reconstruction process.
The experimental results demonstrate that the proposed framework achieves significant energy savings while maintaining high-quality 3D image reconstruction, making it suitable for resource-constrained systems.
Recommended citation: Wu, Y. (2022). Energy Efficient Approximate 3D Image Reconstruction. *IEEE Transactions on Emerging Topics in Computing*, 10(2), 1234–1245. https://doi.org/10.1109/TETC.2022.09559866
Download Paper | Download Slides | Download Bibtex