## paper review: End-to-End Detection with Transformers (a.k.a DETR)

arxiv link key points introduces transformer into the domain of object detection compared to existing approaches, no need for anchor priors and post processing such as NMS to handle with multiple box predictions matching to single gt. this is possible Read more…

## various properties of cairo text_extent

when using cairo with python, the text_extent function call is powerful because it returns coordinate information of the text the user wants to print. However, the information depends on the Context which will be used when calling this function and Read more…

## python operator precedence

https://docs.python.org/3/reference/expressions.html#operator-precedence

## allowing utf8 characters in python csv writer

reference: https://stackoverflow.com/questions/46551955/python-3-csv-utf-8-encoding do this at the start of csv writer to ad BOM

## teacher forcing in training sequence output models

nice summary: https://towardsdatascience.com/what-is-teacher-forcing-3da6217fed1c

## “Stacked Hourglass Networks for Human Pose Estimation” paper review

paper link submitted in 2016 I’m only interested in the stacked hourglass architecture, not about pose segmentation performance. So the points listed below are only related to stacked hourglass architecture. “encoding-decoding” or “conv-deconv” structure is already introduced. This paper goes Read more…

## FCN, UNet, FPN comparison

The three all seem to have “downsampling and then upsampling” idea at the core. But what are the differences? Which one is the correct one to coin when referencing the “downsampling + upsampling” idea? Fully Convolutional Network(FCN) submitted: 2014.11.14https://arxiv.org/pdf/1411.4038.pdf remove Read more…