Related Experiment Video
Updated: Jan 24, 2026
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
1.0K
Differential convolutional neural network.
M Sarıgül1, B M Ozyildirim2, M Avci3
1Iskenderun Technical University, Computer Engineering Department, Hatay, Turkey.
Summary
A new Differential Convolution technique enhances deep learning models by analyzing feature map changes, boosting accuracy by up to 55.29% and improving performance on datasets like ImageNet.
Area of Science:
- Computer Science
- Artificial Intelligence
- Machine Learning
Background:
- Convolutional Neural Networks (CNNs) are popular for deep structure representation.
- Existing CNNs benefit from convolutional layers, driving innovation in convolutional techniques.
Purpose of the Study:
- To introduce a novel convolution technique, Differential Convolution, and an improved back-propagation algorithm.
- To enhance feature map transfer by considering directional activation differences and neighborhood errors.
Main Methods:
- Developed Differential Convolution, adapting mathematical differentiation into the convolutional process.
- Implemented an updated back-propagation algorithm considering neighborhood activation errors.
- Evaluated the technique across four experiment sets on various CNN architectures and datasets.
Main Results:
- Differential Convolution boosted traditional CNN test accuracy by up to 55.29%.
- Achieved notable accuracy improvements on AlexNet with ImageNet dataset (top1: +5.3%, top5: +4.75%).
- Differential VGGNet and NIN models showed high accuracy on CIFAR10 (93.58%, 92.44%) and CIFAR100 (75.06%, 72.65%).
Conclusions:
- Differential Convolution outperforms traditional and other compared convolution techniques.
- The technique is easily adaptable to diverse CNN structures, enhancing popular deep learning models.
- This method offers improved classification performance without increasing the number of filters.
Related Concept Videos
01:17Convolution Properties II
583
The important convolution properties include width, area, differentiation, and integration properties.
The width property indicates that if the durations of input signals are T1 and T2, then the width of the output response equals the sum of both durations, irrespective of the shapes of the two functions. For instance, convolving two rectangular pulses with durations of 2 seconds and 1 second results in a function with a width of 3 seconds.
The area property asserts that the area under the...
The width property indicates that if the durations of input signals are T1 and T2, then the width of the output response equals the sum of both durations, irrespective of the shapes of the two functions. For instance, convolving two rectangular pulses with durations of 2 seconds and 1 second results in a function with a width of 3 seconds.
The area property asserts that the area under the...
583
01:20Convolution Properties I
568
Convolution computations can be simplified by utilizing their inherent properties.
The commutative property reveals that the input and the impulse response of an LTI (Linear Time-Invariant) system can be interchanged without affecting the output:
The commutative property reveals that the input and the impulse response of an LTI (Linear Time-Invariant) system can be interchanged without affecting the output:
568
02:26Protein Networks
4.5K
An organism can have thousands of different proteins, and these proteins must cooperate to ensure the health of an organism. Proteins bind to other proteins and form complexes to carry out their functions. Many proteins interact with multiple other proteins creating a complex network of protein interactions.
These interactions can be represented through maps depicting protein-protein interaction networks, represented as nodes and edges. Nodes are circles that are representative of a protein,...
These interactions can be represented through maps depicting protein-protein interaction networks, represented as nodes and edges. Nodes are circles that are representative of a protein,...
4.5K
02:26Protein Networks
2.8K
2.8K
02:18Network Covalent Solids
16.1K
Network covalent solids contain a three-dimensional network of covalently bonded atoms as found in the crystal structures of nonmetals like diamond, graphite, silicon, and some covalent compounds, such as silicon dioxide (sand) and silicon carbide (carborundum, the abrasive on sandpaper). Many minerals have networks of covalent bonds.
To break or to melt a covalent network solid, covalent bonds must be broken. Because covalent bonds are relatively strong, covalent network solids are typically...
To break or to melt a covalent network solid, covalent bonds must be broken. Because covalent bonds are relatively strong, covalent network solids are typically...
16.1K
01:24Convolution: Math, Graphics, and Discrete Signals
894
In any LTI (Linear Time-Invariant) system, the convolution of two signals is denoted using a convolution operator, assuming all initial conditions are zero. The convolution integral can be divided into two parts: the zero-input or natural response and the zero-state or forced response, with t0 indicating the initial time.
To simplify the convolution integral, it is assumed that both the input signal and impulse response are zero for negative time values. The graphical convolution process...
To simplify the convolution integral, it is assumed that both the input signal and impulse response are zero for negative time values. The graphical convolution process...
894