1. Image recognition tasks can be divided into three levels, according to the abstraction of the processing content, from low to high?

2. The request/response of HUAWEI CAI API in accordance with the design of RESTful API includes the following.

3. Use a 3*3 convolution kernel to convolve a three-channel color image.
What is the number of parameters to be learned?

4. Markov chain is a discrete event stochastic process with Markov properties in exponential science.

5. The accuracy of the target detection position can be evaluated by the IOD value. The iout calculation method is to find the ratio of the union and intersection of the true annotations of the detection results.

6. Which of the following templates can achieve the sharpness of the image?

7. Modifying the H component of the HSV color space will change the image? (Single choice)

8. When taking light, the face is relatively dark. Use gamma correction to enhance the image to achieve a clear light effect.
Which of the Y limit can be selected. (Single choice)

9. Which of the following is pretreatment?

10. What is the correct description about the Gaussian distribution curve?

11. Which of the following options is not a commonly used algorithm for keyword extraction? (Single choice)

12. When photographing a document with a mobile phone, the document in the image is distorted due to the shooting angle. You can find () coordinates of the corresponding key points, () transform the image, and correct the document image.

13. Which of the following options is the method of automatically calculating the gradient.

14. In natural language processing, the most commonly used deep learning algorithm is CNN, because of its memory function?

15. There are a lot of small noises in a photo during storage.
Which operation is the best for noise reduction?

16. What is the difficulty of speech recognition?

17. In the N-gram model, N refers to ignoring the effect of the above words with a distance greater than N. The larger the N, the richer the sequence information contained in the model, and the amount of calculation increases.

18. Which network structure does the convolutional neural network include?

19. In image processing tasks, what is the true label of the data set?

20. Which of the following options is not a method in the development of natural language processing?

21. Image target detection algorithm needs to be completed?

22. The status sequence of HMM can neither be obtained directly nor through observation sequence.

23. Which of the following steps in the MFC extraction process?

24. A model in which several linear models are mixed together is called a linear mixed model.

25. An APP that can display plant names and breeding methods after taking photos of plants in pots. Its core is the use of image classification technology.