Xception

Xception is a convolutional neural network (CNN) architecture widely used in deep learning for image classification. Proposed by Google in 2017, the name "Xception" stands for "Extreme Inception" and was developed based on the Inception architecture. The Xception architecture is particularly notable for its reliance on the principle of depthwise separable convolutions, which enables higher performance with fewer parameters.

Xception improves the information processing pipeline by replacing standard convolutional layers with depthwise separable convolutions, making the structure more modular and efficient. This design is based on the principle of filtering each channel separately and then combining them.

^{Xception Mimarisi (Kaynak =}

Depthwise separable convolutions consist of a two-step process:

1. Depthwise Convolution: Each channel is filtered individually using its own filter. This step performs spatial filtering.
2. Pointwise Convolution: Inter-channel interactions are established through 1x1 convolutions, which rearrange the channel dimensions.

These two steps simulate the operations of a classic convolution more efficiently.

Xception consists of 36 convolutional layers, organized into three main blocks:

• Entry Flow: Transforms the input into low-dimensional representations. It begins with two standard convolutional layers followed by three depthwise separable convolution blocks. During this stage, basic edge and texture features are extracted from the input image while spatial dimensions are reduced to lower computational load for subsequent stages.
• Middle Flow: This is the core learning component of the Xception architecture. It contains eight repeated units, each consisting of three depthwise separable convolutions and a residual connection. These repetitions enable the model to learn more complex and abstract features. The repetitive structure of the middle flow enhances model depth and facilitates the generation of stronger representations.
• Exit Flow: This section transforms the learned features into more concentrated representations suitable for classification. Depthwise separable convolutions and residual connections are also used here. After this block, a global average pooling layer is typically followed by a fully connected layer.

^{(a) Standart CNN (b) Derinlik Yönün Ayrılabilir CNN (Kaynak =}

The Xception architecture achieves more efficient feature extraction compared to classic convolutions by leveraging depthwise separable convolutions.

The Xception architecture has been successfully applied to various image classification and object detection tasks, notably on ImageNet. Similar separable convolution principles have also been adopted in models such as MobileNetV2 and EfficientNet. Its primary application areas include:

• Image classification
• Medical image analysis
• Video analysis
• Industrial quality control
• Transfer learning based models