探索深层卷积神经网络中纹理特征的作用:来自Portilla-Simoncelli统计的见解
Yusuke Hamano1, Shoko Nagasaka2, Hayaru Shouno2
1NEC Corporation, Shiba 5-7-1, Minato-ku, Tokyo, Japan.
概括
深度卷积神经网络 (DCNNs) 使用纹理信息进行图像识别. 本研究使用简化的Portilla-Simoncelli统计 (minPS) 来展示VGG网络如何表示纹理,发现早期层利用minPS特征.
科学领域:
- 计算机视觉 计算机视觉
- 机器学习 机器学习
- 图像识别 图像识别
背景情况:
- 深度卷积神经网络 (DCNNs) 的性能依赖于形状和纹理.
- 了解DCNN的内部表示,特别是纹理编码,仍然是一个挑战.
研究的目的:
- 研究如何在预先训练的VGG网络中表示纹理信息.
- 在此分析中使用 Portilla-Simoncelli 统计数据 (minPS) 的简化版本.
主要方法:
- 从纹理图像中提取了minPS特征.
- 在VGG网络层激活上进行了稀疏回归.
- 使用minPS功能比较原始和VGG合成的图像.
主要成果:
- 早期到中期VGG层的通道被minPS特征描述得很好.
- minPS子组的解释能力在网络层之间发生变化.
- 线性交叉尺度 (LCS) 和能量交叉尺度 (ECS) 的子组显示出有限的解释能力.
结论:
- VGG网络使用特定的minPS特征来表示纹理,特别是在早期的层中.
- 网络的内部纹理表示随着网络深度而演变.
- 合成图像中某些minPS特征的缺失表明VGG没有使用它们.
相关概念视频
Shape and Texture of Coarse Aggregate
230
Aggregate shape is classified based on the relative sharpness or roundness of the edges and corners. This classification includes categories like rounded, angular, elongated, and flaky, each with specific characteristics. Rounded aggregates, fully shaped by attrition, are typical of river or seashore gravel, while angular aggregates, such as crushed rock, have well-defined edges. Aggregates that are elongated and flaky are less desirable, as they can reduce the workability and strength of...
230
Convolution Properties II
224
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...
224
Convolution Properties I
170
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:
170
Convolution: Math, Graphics, and Discrete Signals
282
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...
282
Deconvolution
180
Deconvolution, also known as inverse filtering, is the process of extracting the impulse response from known input and output signals. This technique is vital in scenarios where the system's characteristics are unknown, and they must be inferred from the observable signals.
Deconvolution involves several mathematical techniques to derive the impulse response. One common approach is polynomial division. In this method, the input and output sequences are treated as coefficients of...
Deconvolution involves several mathematical techniques to derive the impulse response. One common approach is polynomial division. In this method, the input and output sequences are treated as coefficients of...
180
Upsampling
254
Managing signal sampling rates is essential in digital signal processing to maintain signal integrity. A decimated signal, characterized by a reduced frequency range due to its lower sampling rate, can be upsampled by inserting zeros between each sample. This upsampling process expands the original spectrum and introduces repeated spectral replicas at intervals dictated by the new Nyquist frequency. To refine this zero-inserted sequence, it is passed through a lowpass filter with a cutoff...
254


