Shulou(Shulou.com)06/02 Report--

Qiu recruit a variety of written interviews, summarize the image processing and C++ topics encountered. What is written down can be remembered, and there should be a lot of people who can't remember. Probably let people who have not experienced know what kind of problems they will encounter, so they can prepare in advance. In addition to the following questions, the most questions are around the projects you have done.

1. Image processing problems

Note that for all the topics that need to be written, the Mat class of OpenCV is not allowed. If the content of the picture needs to be read with a pointer.

Given 0-1 matrix, find the connected domain. (I have encountered N times, there are all written interviews, it is best to be able to use hack code or pseudo code with bare hands. )

Write a function to find the histogram of the grayscale image.

Write a mean filter (median filter).

Write out the Gauss operator, Sobel operator, Laplace operator and so on, as well as the difference in their gradient direction.

Commonly used feature extraction methods.

Commonly used target detection methods.

Commonly used edge extraction methods.

Commonly used interpolation methods.

Commonly used image segmentation algorithms.

Write an image resize function (zoom in and out).

What is the difference between color image, grayscale image, binary image and index image? (what on earth is an index image? )

Common methods of target detection in deep learning, similarities and differences.

The resolution is calculated for a given camera range and image size.

How to detect the car in the picture and identify the car model? what if there is occlusion?

The process of digital recognition.

Introduce neural networks, SVM, AdaBoost, kNN... (each of which can ask for all kinds of details)

Write the gradient descent code.

The difference between convolution neural network and neural network.

The role of convolution layer, the role of pooling layer, the role of fully connected layer.

What are over-fitting and under-fitting respectively and how to improve them.

The function of 1x1 convolution sum.

Calculate the number of parameters in a layer of convolution neural network.

.

II. Interview questions for C++ written examination

This is too much, just say a few more impressive (can still remember). It should be noted that the written interview is generally not a strict distinction between C and C++.

Pointer-related issues, such as:

* what does it mean by being different from the const position?

The difference between a pointer and a reference.

Deep copy and shallow copy.

Pointer array and array pointer.

Pointer function and function pointer.

Wild pointer.

.

Macro.

Const action.

Static action.

Problems related to sizeof.

Realize strcpy function and memcpy function.

Static members (variables and functions) of the class.

Issues related to inheritance.

Vector implementation.

Quick sorting implementation and complexity problems, using scenario problems. (it may also be other common sorting algorithms).

Binary search implementation (recursive non-recursive).

Binary tree traversal (preorder, intermediate order, subsequent, recursive non-recursive).

Problems related to binary tree. (binary tree questions are the most common, which often occur when you fill in the blanks in a variety of choice questions, programming questions, and writing blank code during the interview.)

Recursive related problems. For example, give a recursive function to ask the return result. Or give a problem to be solved recursively.

.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.