When studying the results of a segmentation algorithm using convolutional
neural networks, one wonders about the reliability and consistency of the
results. This leads to questioning the possibility of using such an algorithm
in applications where there is little room for doubt. We propose in this paper
a new attention gate based on the use of Chan-Vese energy minimization to
control more precisely the segmentation masks given by a standard CNN
architecture such as the U-Net model. This mechanism allows to obtain a
constraint on the segmentation based on the resolution of a PDE. The study of
the results allows us to observe the spatial information retained by the neural
network on the region of interest and obtains competitive results on the binary
segmentation. We illustrate the efficiency of this approach for medical image
segmentation on a database of MRI brain images.

本文提出基于 Chan-Vese 能量最小化的新关注门机制，用于更精确地控制标准 CNN 结构（如 U-Net 模型）产生的分割掩模，研究结果显示其在医学图像分割方面具有优越性。

Chan-Vese 注意力 U-Net：用于稳健分割的注意力机制

Chan-Vese Attention U-Net: An attention mechanism for robust  segmentation

Current architectures for multi-modality tasks such as visual question
answering suffer from their high complexity. As a result, these architectures
are difficult to train and require high computational resources. To address
these problems we present a CLIP-based architecture that does not require any
fine-tuning of the feature extractors. A simple linear classifier is used on
the concatenated features of the image and text encoder. During training an
auxiliary loss is added which operates on the answer types. The resulting
classification is then used as an attention gate on the answer class selection.
On the VizWiz 2022 Visual Question Answering Challenge we achieve 60.15 %
accuracy on Task 1: Predict Answer to a Visual Question and AP score of 83.78 %
on Task 2: Predict Answerability of a Visual Question.

该研究提出了一种基于 CLIP 的体系结构，通过将图像和文本编码器的拼接特征上应用简单的线性分类器，并在训练期间添加一项辅助损失，以操作答案类型，并将其作为答案类选择的注意力门，成功解决了多模态任务体系结构高复杂度、难以训练、计算资源要求高的问题，在 VizWiz 2022 视觉问答挑战赛上取得了 60.15％的准确率和 83.78％的平均精度分数。