**Inception v2** is the second generation of Inception convolutional neural network architectures which notably uses [batch normalization](https://paperswithcode.com/method/batch-normalization). Other changes include dropping [dropout](https://paperswithcode.com/method/dropout) and removing [local response normalization](https://paperswithcode.com/method/local-response-normalization), due to the benefits of batch normalization.

**Temporal ROI Align** is an operator for extracting features from other frames' feature maps for current frame proposals by utilizing feature similarity. Considering the features of the same object instance are highly similar among frames in a video, the proposed operator implicitly extracts the most similar ROI features from support frames feature map for target frame proposals based on feature similarity.

Temporal ROIAlign

Temporal RoI Align for Video Object Recognition

Inception v2

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

The goal of **Triplet loss**, in the context of Siamese Networks, is to maximize the joint probability among all score-pairs i.e. the product of all probabilities. By using its negative logarithm, we can get the loss formulation as follows:

$$
L\_{t}\left(\mathcal{V}\_{p}, \mathcal{V}\_{n}\right)=-\frac{1}{M N} \sum\_{i}^{M} \sum\_{j}^{N} \log \operatorname{prob}\left(v p\_{i}, v n\_{j}\right)
$$

where the balance weight $1/MN$ is used to keep the loss with the same scale for different number of instance sets.

Source	Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Year	2000
Data Source	CC BY-SA - https://paperswithcode.com