Nanjing University of Information Science & Technology

Nanjing Horizon Robotics Integrated Circuit Co., Ltd.

Wen Dai

Guojie Wang

Bo Wang

Ruibo Qiu

Kai Chen

Yushu Gao

Xiaodong Zheng

Shuqian Qu

Aili Liu

Kaiwen Kong

Jiaxing Fan

Disclosed are a method and apparatus for accelerating inference of a neural network model, an electronic device, and a medium. The method includes: acquiring image training data, text training data, or speech training data; determining a first neural network model to be accelerated; converting a preset operation on a preset network layer in the first neural network model to a first operation for simulating operational logic of a target operation to obtain a second neural network model; performing, based on the image training data, the text training data, or the speech training data, quantization aware training on the second neural network model by a preset bit width to obtain a third neural network model which is quantized; and converting the first operation of the third neural network model to the target operation, to obtain a target neural network model, which is accelerated, corresponding to the first neural network model.

Method and apparatus for accelerating inference of neural network model, electronic device, and medium

A photogrammetric ground control point quality evaluation method based on Monte Carlo test relates to the field of photogrammetry technology. Firstly, aerial photographs and ground control points from a survey area are obtained, the ground control points are numbered, the aerial photographs are performed with point-placing and aerial triangulation densification. Secondly, a Monte Carlo test experiment is designed, a certain number of the ground control points are selected as control points with the rest as check points, ensuring each ground control point as the control point a certain number of times, and average errors of the ground control points are calculated. Thirdly, average values of the average errors of the ground control points are calculated, standard deviations of the average errors of the ground control points are calculated. Finally, a quality coefficient Q of each ground control point is calculated and evaluated according to quality evaluation standards.

Photogrammetric ground control point quality evaluation method based on Monte Carlo test

A sand dune extraction method based on fusion of positive-negative terrains and light shading includes: obtaining an original digital elevation model of a study region; using a neighborhood analysis method to process the original digital elevation model of the study region, thereby extracting a positive terrain region and a negative terrain region of the study region; using a lighting shading method to perform six directional lighting simulation calculations on each pixel of the original digital elevation model of the study region to obtain a lighting shading map, and extracting a lighting region and a shadow region of the study region from the lighting shading map; using an embedding method to fuse the positive terrain region and the negative terrain region with the lighting region and the shadow region. The sand dune extraction method is a method for remote sensing monitoring and governance of desertification regions.

Sand dune extraction method based on fusion of positive-negative terrains and light shading

Growing community of inventors

Nanjing, China

Wen Dai