Deconvolution Calculation Using Reused Convolution Module
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic apparatuses face challenges in minimizing the area occupied by hardware for both convolution and deconvolution calculations, leading to inefficient use of space due to the need for separate hardware for each operation.
Innovation Solution
An electronic apparatus and method that utilizes a convolution calculation process to perform deconvolution calculations by obtaining calculation information, determining output data size, generating memory addresses, performing convolution calculations, storing intermediate values, and accumulating them to obtain deconvolution results, thereby eliminating the need for dedicated deconvolution hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If separate hardware is configured for convolution calculation and deconvolution calculation, then both operations can be performed independently, but the hardware area occupied becomes excessively large
Solution Approach 1:
The patent makes the convolution calculation module universal by enabling it to perform both convolution calculation and deconvolution calculation through software control. The same hardware resources (processors, memory, address generation module) are reused for both operations, eliminating the need for separate dedicated hardware for deconvolution while maintaining independent operation capability through software-based deconvolution implementation
2Adaptability or versatility
If dedicated deconvolution hardware is added to an electronic apparatus with convolution hardware, then deconvolution operations can be performed, but the overall hardware area increases significantly
Solution Approach 1:
The convolution calculation module is designed to be multi-functional, capable of performing both convolution and deconvolution operations. The control module configures the same hardware resources differently based on the operation type, allowing the system to adapt to different computational needs without expanding hardware area
Solution Approach 2:
The system dynamically reconfigures the convolution calculation module's operation mode based on whether convolution or deconvolution is required. The control module adjusts parameters and data flow dynamically, enabling the same hardware to adapt its function in real-time without physical reconfiguration
Data Source
AI summary
Disclosed is an electronic apparatus including a processor configured to obtain calculation information based on input data of a deconvolution calculation being input, obtain a size of output data based on the obtained calculation information, obtain a plurality of memory address values corresponding to a size of the output data using an address generation module, perform convolution calculation based on the calculation information using a convolution calculation module to obtain an intermediate value in the convolution calculation process, obtain a memory address value corresponding to the obtained intermediate value of the plurality of obtained memory addresses using the address generation module, store the obtained intermediate value in the memory address value corresponding to the intermediate value, accumulate at least one intermediate value based on the memory address value corresponding to the intermediate value using a cumulative calculation module, and obtain a deconvolution calculation value with respect to the input data based on the accumulated at least one intermediate value.