Imaging device and focus adjustment method

Patent metadata

Field	Value
Publication number	US-2019387175-A1
Application number	US-201916432738-A
Country	US
Kind code	A1
Filing date	Jun 5, 2019
Priority date	Jun 14, 2018
Publication date	Dec 19, 2019
Grant date	—

Abstract

An imaging device, comprising an image sensor that receives subject light and generates image data, and a processor comprising a focus control section, an index generating section, and a control section, wherein the focus control section performs focus detection based on the image data, and controls focus drive based on focus detection results, the index generating section is input with the image data, and generates a first index representing which image of a given plurality of types of image the image data is close to, and a second index representing Bokeh state of an image corresponding to the image data, and the control section changes control of focus drive by the focus control section based on output of the index generating section.

First claim

What is claimed is:

1 . An imaging device, comprising:
an image sensor that receives subject light and generates image data; and a processor comprising a focus control section, an index generating section and a control section, wherein the focus control section performs focus detection based on the image data, and controls focus drive based on focus detection results; the index generating section is input with the image data, and generates a first index representing which image of a given plurality of types of image the image data is close to, and a second index representing Bokeh state of an image corresponding to the image data; and the control section changes control of focus drive by the focus control section based on output of the index generating section.

2 . The imaging device of claim 1 , wherein:
the index generating section has a first neural network that generates an index representing which of a plurality of image types the image data is, and a second neural network that generates an index representing Bokeh state of an image corresponding to the image data.

3 . The imaging device of claim 2 , wherein:
the index generating section divides image data into a plurality of image data in accordance with regions of the image, and respective image data that has been divided is input to the first and/or second neural network.

4 . The imaging device of claim 1 , wherein:
the index generating section has a neural network that includes convolution layers at an initial stage.

5 . The imaging device of claim 1 , wherein:
the index generating section includes night scene, or periodicity-containing subject scene, or a scene with a mix of near and far objects, or scenes other than these, as a plurality of image classifications.

6 . The imaging device of claim 5 , wherein:
the focus control section performs correction of an evaluation value generated as a result of focus detection in accordance with brightness, in a case where an index output by the index generating section represents a night scene.

7 . The imaging device of claim 5 , wherein:
the focus control section performs focus detection by changing focus to a close-up end, in a case where an index output by the index generating section represents the scene with a mix of near and far objects.

8 . The imaging device of claim 5 , wherein:
in a case where an index that is output by the index generating section represents a night scene, and an index representing Bokeh state represents a larger Bokeh amount than the specified amount, the focus control section makes a threshold value, for determining in focus, larger.

9 . The imaging device of claim 2 , wherein:
the second neural network of the index generating section has a plurality of neural networks in accordance with image classifications, and generates the second index by selecting the plurality of neural networks based on the first index.

10 . The imaging device of claim 2 , wherein:
the index generating section divides image data into a plurality of image data in accordance with image region, performs image classification using the first neural network for respective regions and all images, and generates the first index based on the image classification results.

11 . The imaging device of claim 1 , wherein:
in a case where the first index represents a significant Bokeh state, the focus control section drives the focus lens at a higher speed than normal.

12 . A focus adjustment method, comprising:
receiving subject light using an image sensor and generating image data; inputting the image data, and generating a first index representing which image of a given plurality of types of image the image data is close to, and a second index representing Bokeh state of an image corresponding to the image data; and performing focus detection based on the image data, and, when controlling a focus operation based on focus detection results, changing the focus drive control based on the first index and the second index.

13 . The focus adjustment method of claim 12 , wherein:
a night scene, or periodicity-containing subject scene, or a scene with a mix of near and far objects, or scenes other than these, are included as the plurality of image classifications.

14 . The focus adjustment method of claim 13 , wherein:
an evaluation value, generated as a result of focus detection, is corrected in accordance with brightness, in a case where the first index represents the night scene.

15 . The focus adjustment method of claim 13 , wherein:
when the first index represents a scene with a mix of near and far objects, focus detection is performed by changing focus to a close-up end.

16 . The focus adjustment method of claim 13 , wherein:
when the first index represents the night scene, and the second index represents a Bokeh amount that is larger than a specified amount, a threshold value for determining focus is made higher.

17 . A non-transitory computer-readable medium storing a processor executable code, which when executed by at least one processor, performs a focus adjusting method, the focus adjusting method comprising:
receiving subject light using an image sensor and generating image data; inputting the image data, and generating a first index representing which image of a given plurality of types of image the image data is close to, and a second index representing Bokeh state of an image corresponding to the image data; and performing focus detection based on the image data, and, when controlling a focus operation based on focus detection results, changing the focus drive control based on the first index and the second index.

18 . The non-transitory computer-readable medium of claim 17 , wherein:
a night scene, or periodicity-containing subject scene, or a scene with a mix of near and far objects, or scenes other than these, are included as the plurality of image classifications.

19 . The non-transitory computer-readable medium of claim 18 , wherein:
an evaluation value, generated as a result of focus detection, is corrected in accordance with brightness, in a case where the first index represents the night scene.

20 . The non-transitory computer-readable medium of claim 18 , wherein:
when the first index represents the scene with a mix of near and far objects, focus detection is performed by changing focus to a close-up end.

Assignees

• Olympus Corp

Inventors

• Kikuchi Tetsuo
• Toyoda Tetsuya

Classifications

• H04N23/667

  Camera operation mode switching, e.g. between still and video, sport and normal or high- and low-resolution modes · CPC title

• H04N23/959

  by adjusting depth of field during image capture, e.g. maximising or setting range based on scene characteristics · CPC title

• H04N23/672 Primary

  based on the phase difference signals · CPC title

• H04N25/704

  Pixels specially adapted for focusing, e.g. phase difference pixel sets · CPC title

• H04N23/675

  comprising setting of focusing regions · CPC title