1/*
2 * This file is part of FFmpeg.
3 *
4 * FFmpeg is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public
6 * License as published by the Free Software Foundation; either
7 * version 2.1 of the License, or (at your option) any later version.
8 *
9 * FFmpeg is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
12 * Lesser General Public License for more details.
13 *
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with FFmpeg; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17 */
18
19#ifndef AVUTIL_DETECTION_BBOX_H
20#define AVUTIL_DETECTION_BBOX_H
21
22#include "rational.h"
23#include "avassert.h"
24#include "frame.h"
25
26typedef struct AVDetectionBBox {
27    /**
28     * Distance in pixels from the left/top edge of the frame,
29     * together with width and height, defining the bounding box.
30     */
31    int x;
32    int y;
33    int w;
34    int h;
35
36#define AV_DETECTION_BBOX_LABEL_NAME_MAX_SIZE 64
37
38    /**
39     * Detect result with confidence
40     */
41    char detect_label[AV_DETECTION_BBOX_LABEL_NAME_MAX_SIZE];
42    AVRational detect_confidence;
43
44    /**
45     * At most 4 classifications based on the detected bounding box.
46     * For example, we can get max 4 different attributes with 4 different
47     * DNN models on one bounding box.
48     * classify_count is zero if no classification.
49     */
50#define AV_NUM_DETECTION_BBOX_CLASSIFY 4
51    uint32_t classify_count;
52    char classify_labels[AV_NUM_DETECTION_BBOX_CLASSIFY][AV_DETECTION_BBOX_LABEL_NAME_MAX_SIZE];
53    AVRational classify_confidences[AV_NUM_DETECTION_BBOX_CLASSIFY];
54} AVDetectionBBox;
55
56typedef struct AVDetectionBBoxHeader {
57    /**
58     * Information about how the bounding box is generated.
59     * for example, the DNN model name.
60     */
61    char source[256];
62
63    /**
64     * Number of bounding boxes in the array.
65     */
66    uint32_t nb_bboxes;
67
68    /**
69     * Offset in bytes from the beginning of this structure at which
70     * the array of bounding boxes starts.
71     */
72    size_t bboxes_offset;
73
74    /**
75     * Size of each bounding box in bytes.
76     */
77    size_t bbox_size;
78} AVDetectionBBoxHeader;
79
80/*
81 * Get the bounding box at the specified {@code idx}. Must be between 0 and nb_bboxes.
82 */
83static av_always_inline AVDetectionBBox *
84av_get_detection_bbox(const AVDetectionBBoxHeader *header, unsigned int idx)
85{
86    av_assert0(idx < header->nb_bboxes);
87    return (AVDetectionBBox *)((uint8_t *)header + header->bboxes_offset +
88                               idx * header->bbox_size);
89}
90
91/**
92 * Allocates memory for AVDetectionBBoxHeader, plus an array of {@code nb_bboxes}
93 * AVDetectionBBox, and initializes the variables.
94 * Can be freed with a normal av_free() call.
95 *
96 * @param out_size if non-NULL, the size in bytes of the resulting data array is
97 * written here.
98 */
99AVDetectionBBoxHeader *av_detection_bbox_alloc(uint32_t nb_bboxes, size_t *out_size);
100
101/**
102 * Allocates memory for AVDetectionBBoxHeader, plus an array of {@code nb_bboxes}
103 * AVDetectionBBox, in the given AVFrame {@code frame} as AVFrameSideData of type
104 * AV_FRAME_DATA_DETECTION_BBOXES and initializes the variables.
105 */
106AVDetectionBBoxHeader *av_detection_bbox_create_side_data(AVFrame *frame, uint32_t nb_bboxes);
107#endif
108