1. Overview The target of face detection is to find out the corresponding positions of all human faces in the image. The output of the algorithm is the coordinates of the external rectangle of the face in the image, and may also include information such as posture, tilt angle, etc.Common face detection databases include FDDB and WIDER FACE.
2, FDDB Official Website: http://vis-www.cs.umass.edu/fddb/
FDDB has a total of 2,845 images, 5,171 images, face unconstrained environment, face difficulty is bigger, there are facial expressions, double chin, lighting changes, wearing, exaggerated hairstyle, occlusion and other difficulties, is the most commonly used database of the target.
1. Compile Computational Geometry Algorithms Library The compilation of this version has a lot of guidance in official documents, but there are still some details that need attention. The general situation is as follows: When compiling demo, using too new qt (such as qt5.8) will cause an interface undefined error.The final version is listed as follows: qt5.6msvc201564, boost 1.60, libQGLViewer2.6.3, tbb64 201702 ..., Eigen3, 0. The official website downloads CGAL-4.11-Setup.exe.
concurrency control In computer science, especially in the fields of program design, operating system, multiprocessor and database, Concurrency control is a mechanism to ensure timely correction of errors caused by concurrent operations. The task of concurrency control in a database management system (DBMS) is to ensure that when multiple transactions access the same data in the database at the same time, the isolation and unity of the transactions and the
In the process of software performance testing, the preparation of test data is a very systematic and heavy work.How to prepare a large amount of test data supporting different business operations and different test types to meet the requirements of load stress testing is an important topic often faced in the performance testing process.China's Software Judgement Center has always attached importance to the preparation of performance test data in the performance test process, thus ensuring the smooth performance test and the accuracy and effectiveness of the performance test results.
brief introduction seetaface is developed by the face recognition research team led by researcher Institute of Computing Technology, Chinese Academy of Sciences Shiguang.The code is implemented based on C++ and does not rely on third-party libraries.However, the current open source code is compiled on windows vs, which is still very troublesome for us mac/linux users.After several days of study, I finally got a comprehensive understanding of seetaface.Next, listen to me.
MODBUS protocol was proposed by Modicon Company in 1979. It is a brand created by Dick Morley, known as the father of PLC.MODBUS is the first bus protocol used in industrial field in the world. It can be said that its appearance marks that industrial field is striding forward from analog era to communication era.
Dick Morley, Father of PLC
Although more than 40 years have passed, MODBUS protocol is still alive and free, which is an important reason.
Source Article: https://github.com/BVLC/caffe/tree/master/examples/imagenet
Due to the secondary development of in-depth learning using CAFFE (Convolutional Architecture for Fast Feature Embedding), I found some operating procedures of how to train the network above. Not every sentence is translated, but it is not original after all. I have the right to use it as my learning notes.
1. Download training and validation data sets from ImageNet;They are stored in the following formats:/path/to/imagenet/train/n0144074/n0144074 _ 10026.
1, cvLoadImage: load image file into memory;
2, cvNamedWindow: Create a window on the screen;
3, cvShowImage: displays the image in a created window;
4, cvWaitKey: Pause the program and wait for the user to trigger a key operation;
5, cvReleaseImage: Free the memory allocated by the image file;
6. cvDestroyWindow: Destroy the window displaying the image file;
7, cvCreateFileCapture: determine the AVI file to be read through parameter setting;
Recently, more and more people asked me how I can quickly become a data scientist. I can earn more and work less. Headhunters chase job-hopping all day, doubling.My general answer is that there is no such thing as a good thing, but engaging in a Kaggle competition will help you become a data scientist quickly, and then it depends on how much you earn.As to what the Kaggle competition is, please Google yourself for space reasons.