View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0002712 | Composr | core | public | 2016-07-14 00:25 | 2021-03-15 18:41 |
| Reporter | Chris Graham | Assigned To | |||
| Severity | Feature-request | ||||
| Status | non-assigned | Resolution | open | ||
| Product Version | |||||
| Fixed in Version | |||||
| Summary | 0002712: Smart thumbnailing based on image content | ||||
| Description | We can integrate some third party code to get thumbnailing based on any of: - Most 'energetic' part of the image (CropEntropy: https://github.com/stojg/crop) - Faces (https://github.com/mauricesvay/php-facedetection [there's also a good Microsoft API and OpenCV PHP extension, but those are excessive - the PHP-based algorithm here is sound and a port of a popular JavaScript one]) We'd implement it as an extension of the 'where' parameter in the convert_image_plus function. Instead of start/end/both we provide a value of either 'energetic' or 'face'. These essentially detect a central point which becomes the crop/pad origin. | ||||
| Tags | No tags attached. | ||||
| Time estimation (hours) | 10 | ||||
| Sponsorship open | |||||
|
|
Further to defining the central point, we may also want to respect face width's/height's for the case of the face being bigger than the thumbnail. We'd need to do some pre-scaling down in the convert_image_plus function, so that what is passed through to convert_image has a face fitting the thumbnail with some spare. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2016-07-14 00:25 | Chris Graham | New Issue | |
| 2016-07-14 11:17 | Chris Graham | Note Added: 0004125 | |
| 2016-07-14 11:17 | Chris Graham | Note Edited: 0004125 | View Revisions |
| 2021-03-15 18:41 | Chris Graham | Summary | Thumbnailing based on content => Smart thumbnailing based on image content |
