CN114821263A - Weak texture target pose estimation method based on feature fusion - Google Patents
Weak texture target pose estimation method based on feature fusion Download PDFInfo
- Publication number
- CN114821263A CN114821263A CN202210623453.4A CN202210623453A CN114821263A CN 114821263 A CN114821263 A CN 114821263A CN 202210623453 A CN202210623453 A CN 202210623453A CN 114821263 A CN114821263 A CN 114821263A
- Authority
- CN
- China
- Prior art keywords
- features
- point
- target
- fusion
- pose estimation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a weak texture target pose estimation method based on feature fusion, which comprises the following steps of firstly, performing semantic segmentation on an RGB image where a target is located to obtain a target pixel mask and a minimum bounding box corresponding to the target; secondly, cutting the RGB image by adopting a minimum bounding box; thirdly, extracting color features by adopting a convolutional neural network; fourthly, acquiring a depth image; fifthly, segmenting the depth image and converting the depth image into point cloud data; sixthly, acquiring point features, local geometric features and global geometric features in the point cloud data; fusing the color features with the point features, the local geometric features and the global geometric features to obtain target fusion features; and eighthly, inputting the target fusion characteristics into a pose estimation network and outputting a pose estimation result. The method can be effectively applied to the pose estimation of the weak texture target, the problem that the pose estimation is insufficient in local feature consideration during feature fusion is solved, the accuracy of the pose estimation is improved, and the method is convenient to popularize and use.
Description
Technical Field
The invention belongs to the technical field of target pose estimation, and particularly relates to a weak texture target pose estimation method based on feature fusion.
Background
The target pose estimation can load a virtual object generated by a computer to a real image sequence, acquire the pose of an object, help a mechanical arm to clamp the object, and have wide application in the fields of mechanical arm grabbing, augmented reality and the like.
In the prior art, a feature descriptor is used to extract target features to realize pose estimation, for example, in a pose estimation method combining a SURF descriptor and a self-encoder (CN114037742A), SURF feature points of a color image are extracted, features extracted by a convolution self-encoder and corresponding pose information in rendering data form an offline feature template, and K feature vectors with the minimum distance in the feature template are selected to vote to obtain a 6D pose of a target. The invention reduces manual labeling, reduces environment complexity and reduces the calculation amount. However, since the use of a feature descriptor such as SURF requires a target to have a rich texture pattern, this method has a problem that it is difficult to extract features from a weak texture target and the effect of the weak texture target is poor. A weak texture object pose estimation method (CN113223181A) is optimized aiming at a weak texture target, and the method obtains color embedding characteristics and geometric embedding characteristics through a color image and a point cloud respectively, extracts a position dependent characteristic image by utilizing a self-attention mechanism, and carries out pose estimation after pixel-by-pixel fusion. The invention can enrich the information of the characteristics of each pixel, adaptively adjust the weights of different characteristics and improve the identification precision of each pixel. However, the method of dense fusion is used, and the influence of local features existing between point clouds on pose estimation is ignored, so that the accuracy of pose estimation is limited.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a weak texture target pose estimation method based on feature fusion aiming at the defects in the prior art, the method has simple steps, reasonable design and convenient implementation, can be effectively applied to the pose estimation of weak texture targets, solves the problem of insufficient consideration of local features during the feature fusion of the pose estimation, improves the accuracy of the pose estimation, has stronger model adaptability and good use effect, and is convenient to popularize and use. .
In order to solve the technical problems, the invention adopts the technical scheme that: a weak texture target pose estimation method based on feature fusion comprises the following steps:
firstly, performing semantic segmentation on an RGB image where a target is located to obtain a target pixel mask and a minimum bounding box corresponding to the target;
step two, cutting the RGB image by adopting the minimum bounding box to obtain the cut RGB image;
extracting color features of the target in the cut RGB image by adopting a convolutional neural network;
step four, obtaining a depth image of the cut RGB image;
fifthly, dividing the depth image by adopting the mask and converting the depth image into point cloud data;
sixthly, acquiring point features, local geometric features and global geometric features in the point cloud data;
step seven, fusing the color features with the point features, the local geometric features and the global geometric features to obtain target fusion features;
and step eight, inputting the target fusion characteristics into a pose estimation network, and outputting a pose estimation result.
In the above weak texture target pose estimation method based on feature fusion, the specific process of extracting the color features of the target in the clipped RGB image by using the convolutional neural network in step three includes:
step 301, adopting 18 convolution layers to carry out down-sampling on the cut RGB image characteristics to obtain characteristics with dimension of 512;
and 302, performing up-sampling on the features by adopting four up-sampling layers to obtain 32-dimensional color features.
In the above weak texture target pose estimation method based on feature fusion, the specific process of acquiring the point features, the local geometric features and the global geometric features in the point cloud data in the sixth step includes:
601, extracting point characteristics in point cloud data by using a PointNet network;
step 602, randomly selecting 256 position points, and fusing the characteristics of the position points to reduce the influence of noise on target shielding and segmentation;
step 603, finding 128 points which are uniformly distributed in the space by adopting a mode of farthest point sampling;
step 604, dividing the sphere with the fixed radius into a local area by taking each uniformly distributed point as a central point;
605, extracting features of point clouds in three spatial scales of 0.05cm, 0.1cm and 0.2cm in each local area by PointNet, and performing connection aggregation to form local geometric features;
step 606, adopting a farthest point sampling mode to find 64 points which are uniformly distributed in the space;
step 607, dividing the sphere with fixed radius into a local area by taking each evenly distributed point as a central point;
step 608, in each local area, performing feature extraction and aggregation on the point clouds in two spatial scales of 0.2cm and 0.3 cm;
and step 609, extracting the global geometric features of the target on the basis of the local geometric features by adopting MLP.
In the above weak texture target pose estimation method based on feature fusion, the specific process of fusing the color features with the point features, the local geometric features and the global geometric features to obtain the target fusion features includes:
701, performing 1d convolution operation on the color features, and fusing the color features and the point features to form point fusion features;
step 702, performing 1d convolution operation on the point fusion features again, and fusing the point fusion features with the local geometric features to form local fusion features;
and step 703, fusing the global geometric feature in step 609, the point fusion feature in step 701 and the local fusion feature in step 702 to form a final target fusion feature.
In the above weak texture target pose estimation method based on feature fusion, the specific process of inputting the target fusion features into the pose estimation network and outputting the pose estimation result in step eight includes:
step 801, taking the target fusion characteristics as a training set, and training a pose estimation network;
step 802, the pose estimation network predicts the rotation and translation of the target and the confidence of pose prediction;
step 803, using the pose prediction made by the position point with the highest confidence coefficient as the initial pose;
and 804, optimizing the initial pose by adopting a four-layer fully-connected network to obtain a final pose estimation result.
In the above weak texture target pose estimation method based on feature fusion, in step eight, the pose estimation network includes a Loss function, the Loss function weights pose Loss by confidence, and the Loss function Loss is:
wherein i represents the ith point of the N position points,
representing the pose loss of the ith point in the N position points, c i And representing the confidence coefficient of the predicted pose at the ith point, wherein omega represents a balance parameter.
Compared with the prior art, the invention has the following advantages:
1. the method has simple steps, reasonable design and convenient realization.
2. According to the method, the target features with higher quality are obtained by paying attention to the geometric features with fine granularity existing between local point cloud data.
3. The method solves the problem that the pose estimation is insufficient in local feature consideration during feature fusion, and improves the accuracy of the pose estimation.
4. The method can be effectively applied to the pose estimation of the weak texture target, and is high in precision, stronger in model adaptability, good in using effect and convenient to popularize and use.
In conclusion, the method provided by the invention has the advantages of simple steps, reasonable design and convenience in implementation, can be effectively applied to pose estimation of the weak texture target, solves the problem that the pose estimation is insufficient in consideration of local features during feature fusion, improves the accuracy of pose estimation, has stronger model adaptability and good use effect, and is convenient to popularize and use.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
FIG. 1 is a flow chart of a method of the present invention;
FIG. 2 is a diagram of a network architecture of the present invention;
FIG. 3 is a diagram of the visualization effect of the estimation result of each target pose according to the present invention.
Detailed Description
As shown in fig. 1 and fig. 2, the weak texture target pose estimation method based on feature fusion of the present invention includes the following steps:
firstly, performing semantic segmentation on an RGB image where a target is located to obtain a target pixel mask and a minimum bounding box corresponding to the target;
step two, cutting the RGB image by adopting the minimum bounding box to obtain the cut RGB image;
extracting color features of the target in the cut RGB image by adopting a convolutional neural network;
step four, obtaining a depth image of the cut RGB image;
fifthly, dividing the depth image by adopting the mask and converting the depth image into point cloud data;
sixthly, acquiring point features, local geometric features and global geometric features in the point cloud data;
step seven, fusing the color features with the point features, the local geometric features and the global geometric features to obtain target fusion features;
and step eight, inputting the target fusion characteristics into a pose estimation network, and outputting a pose estimation result.
In this embodiment, the specific process of extracting the color feature of the target in the clipped RGB image by using the convolutional neural network in step three includes:
step 301, adopting 18 convolution layers to carry out down-sampling on the cut RGB image characteristics to obtain characteristics with dimension of 512;
and 302, performing up-sampling on the features by adopting four up-sampling layers to obtain 32-dimensional color features.
In this embodiment, the specific process of acquiring the point feature, the local geometric feature and the global geometric feature in the point cloud data in the sixth step includes:
601, extracting point characteristics in point cloud data by using a PointNet network;
step 602, randomly selecting 256 position points, and fusing the characteristics of the position points to reduce the influence of noise on target shielding and segmentation;
step 603, finding 128 points which are uniformly distributed in the space by adopting a farthest point sampling mode;
step 604, dividing the sphere with the fixed radius into a local area by taking each uniformly distributed point as a central point;
605, extracting features of point clouds in three spatial scales of 0.05cm, 0.1cm and 0.2cm in each local area by PointNet, and performing connection aggregation to form local geometric features;
step 606, adopting a farthest point sampling mode to find 64 points which are uniformly distributed in the space;
step 607, dividing the sphere with fixed radius into a local area by taking each evenly distributed point as a central point;
step 608, in each local area, performing feature extraction and aggregation on the point clouds in two spatial scales of 0.2cm and 0.3 cm;
and step 609, extracting the global geometric features of the target on the basis of the local geometric features by adopting MLP.
In this embodiment, the specific process of fusing the color feature with the point feature, the local geometric feature, and the global geometric feature to obtain the target fusion feature in the seventh step includes:
701, performing 1d convolution operation on the color features, and fusing the color features and the point features to form point fusion features;
step 702, performing 1d convolution operation on the point fusion features again, and fusing the point fusion features with the local geometric features to form local fusion features;
and step 703, fusing the global geometric feature in step 609, the point fusion feature in step 701 and the local fusion feature in step 702 to form a final target fusion feature.
In this embodiment, the step eight of inputting the target fusion feature into the pose estimation network, and the specific process of outputting the pose estimation result includes:
step 801, taking the target fusion characteristics as a training set, and training a pose estimation network;
step 802, the pose estimation network predicts the rotation and translation of the target and the confidence of pose prediction;
step 803, using the pose prediction made by the position point with the highest confidence coefficient as the initial pose;
and 804, optimizing the initial pose by adopting a four-layer fully-connected network to obtain a final pose estimation result.
In this embodiment, the pose estimation network in the eighth step includes a Loss function, where the Loss function weights pose Loss by confidence, and the Loss function Loss is:
wherein i represents the ith point of the N position points,
representing the pose loss of the ith point in the N position points, c i And representing the confidence coefficient of the predicted pose at the ith point, wherein omega represents a balance parameter.
In the specific implementation process, the first-stage reactor,
coordinates of target 3D model sampling points are respectively processed through a Ground Truth pose matrix [ R | t ]]And estimating a pose matrix
After transformation, the average distance between the corresponding point coordinates.
The calculation formula is as follows:
wherein M represents a target 3D model sampling point set, i represents the ith point of N position points, p is used as a mark and represents the position and pose loss, and x j J-th point representing a set of sample points, (Rx) j + t) represents the coordinates of the sampling point after being transformed by the Ground Truth pose,
and representing the coordinates after pose transformation by the first point estimation.
For an object with rotational symmetry, the pose loss is defined as: coordinates of sampling points of the target 3D model are obtained by estimating a pose matrix
Transformed coordinatesAnd through the Ground Truth pose matrix [ R | t]After transformation, the average distance between the coordinates of the closest points is calculated by the formula:
wherein x is k Is represented by the formula j The closest point.
The invention is based on an Intel (R) Xeon (R) CPU E5-2678 v32.50GHz processor and an NVIDIA GeForce RTX 2080 display card, and is developed by using Pytroch 0.4.1 and Python3.6 and performing network accelerated training by using CUDA9.0 and cuDNN7.6.4 under an Ubuntu16.04 system.
Table 1 shows the initial pose results and pose optimization results of the method in LineMOD datasets. The data set contains sequences of weak texture objects in 13 complex backgrounds, each sequence containing 1100 and 1300 RGB-D images. Ape, Benchvise, Camera, Can, Cat, Driller, Duck, Eggbox, Glue, Holepuncher, Iron, Lamp, and Phone, respectively. The Eggbox and the Glue targets have rotational symmetry, and the sizes of the targets are different. The experiment divides 15% of the RGB-D images for each target into a training set, with the remainder being the test set. The training set contains 2372 total RGB-D images for the 13 targets and the test set contains 13406 total RGB-D images for the 13 targets.
The comparison index adopts ADD precision, and an ADD calculation formula is as follows:
wherein, Num pre Number of correct pose estimates, Num GT Representing the number of all real poses. And if the ADD is less than 10% of the target maximum diameter value, the pose estimation is considered to be correct, and the ADD calculation formula is as follows:
the higher the accuracy is, the better the pose estimation method is.
TABLE 1 pose optimization results
When the method extracts the local geometric characteristics of the point cloud, the point cloud characteristics of the point clouds in different radius spaces are respectively extracted in a local area, and because different targets are trained simultaneously, the selected multi-scale radii have the same size. The local geometric feature extraction is not fine enough for the target with a small size, and therefore the experimental precision is affected. And for the targets with large and medium body types, the pose estimation precision is better.
In order to further verify the effect of the method, the pose estimation result of the target is visualized, and the result graph is shown in fig. 3.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and all simple modifications, changes and equivalent structural changes made to the above embodiment according to the technical spirit of the present invention still fall within the protection scope of the technical solution of the present invention.
Claims (6)
1. A weak texture target pose estimation method based on feature fusion is characterized by comprising the following steps:
firstly, performing semantic segmentation on an RGB image where a target is located to obtain a target pixel mask and a minimum bounding box corresponding to the target;
step two, cutting the RGB image by adopting the minimum bounding box to obtain the cut RGB image;
extracting color features of the target in the cut RGB image by adopting a convolutional neural network;
step four, obtaining a depth image of the cut RGB image;
fifthly, dividing the depth image by adopting the mask and converting the depth image into point cloud data;
sixthly, acquiring point features, local geometric features and global geometric features in the point cloud data;
step seven, fusing the color features with the point features, the local geometric features and the global geometric features to obtain target fusion features;
and step eight, inputting the target fusion characteristics into a pose estimation network, and outputting a pose estimation result.
2. The feature fusion-based weak texture object pose estimation method according to claim 1, wherein the specific process of extracting the color features of the objects in the clipped RGB image by using the convolutional neural network in the third step comprises:
step 301, adopting 18 convolution layers to carry out down-sampling on the cut RGB image characteristics to obtain characteristics with dimension of 512;
and 302, performing up-sampling on the features by adopting four up-sampling layers to obtain 32-dimensional color features.
3. The weak texture target pose estimation method based on feature fusion as claimed in claim 1, wherein the specific process of acquiring the point features, the local geometric features and the global geometric features in the point cloud data in the sixth step comprises:
601, extracting point characteristics in point cloud data by using a PointNet network;
step 602, randomly selecting 256 position points, and fusing the characteristics of the position points to reduce the influence of noise on target shielding and segmentation;
step 603, finding 128 points which are uniformly distributed in the space by adopting a farthest point sampling mode;
step 604, dividing the sphere with the fixed radius into a local area by taking each uniformly distributed point as a central point;
605, extracting features of point clouds in three spatial scales of 0.05cm, 0.1cm and 0.2cm in each local area by PointNet, and performing connection aggregation to form local geometric features;
step 606, adopting a farthest point sampling mode to find 64 points which are uniformly distributed in the space;
step 607, dividing the sphere with fixed radius into a local area by taking each evenly distributed point as a central point;
step 608, in each local area, performing feature extraction and aggregation on point clouds in two spatial scales of 0.2cm and 0.3 cm;
and step 609, extracting the global geometric features of the target on the basis of the local geometric features by adopting MLP.
4. The weak texture target pose estimation method based on feature fusion as claimed in claim 3, wherein the specific process of fusing the color feature with the point feature, the local geometric feature and the global geometric feature to obtain the target fusion feature in the seventh step includes:
701, performing 1d convolution operation on the color features, and fusing the color features and the point features to form point fusion features;
step 702, performing 1d convolution operation on the point fusion features again, and fusing the point fusion features with the local geometric features to form local fusion features;
and step 703, fusing the global geometric feature in step 609, the point fusion feature in step 701 and the local fusion feature in step 702 to form a final target fusion feature.
5. The weak texture target pose estimation method based on feature fusion as claimed in claim 4, wherein the specific process of inputting the target fusion features into the pose estimation network and outputting the pose estimation result in step eight includes:
step 801, taking the target fusion characteristics as a training set, and training a pose estimation network;
step 802, the pose estimation network predicts the rotation and translation of the target and the confidence of pose prediction;
step 803, using the pose prediction made by the position point with the highest confidence coefficient as the initial pose;
and 804, optimizing the initial pose by adopting a four-layer fully-connected network to obtain a final pose estimation result.
6. The weak texture target pose estimation method based on feature fusion of claim 5, wherein the pose estimation network in the eighth step comprises a Loss function, the Loss function weights pose Loss through confidence, and the Loss function Loss is:
wherein i represents the ith point of the N position points,
representing the pose loss of the ith point in the N position points, c i And representing the confidence coefficient of the predicted pose of the ith point, wherein omega represents a balance parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210623453.4A CN114821263A (en) | 2022-06-01 | 2022-06-01 | Weak texture target pose estimation method based on feature fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210623453.4A CN114821263A (en) | 2022-06-01 | 2022-06-01 | Weak texture target pose estimation method based on feature fusion |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114821263A true CN114821263A (en) | 2022-07-29 |
Family
ID=82520157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210623453.4A Pending CN114821263A (en) | 2022-06-01 | 2022-06-01 | Weak texture target pose estimation method based on feature fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114821263A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115331263A (en) * | 2022-09-19 | 2022-11-11 | 北京航空航天大学 | Robust attitude estimation method and application thereof in orientation judgment and related method |
| CN115661929A (en) * | 2022-10-28 | 2023-01-31 | 北京此刻启动科技有限公司 | Time sequence feature coding method and device, electronic equipment and storage medium |
| CN115880717A (en) * | 2022-10-28 | 2023-03-31 | 北京此刻启动科技有限公司 | Heatmap key point prediction method and device, electronic equipment and storage medium |
-
2022
- 2022-06-01 CN CN202210623453.4A patent/CN114821263A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115331263A (en) * | 2022-09-19 | 2022-11-11 | 北京航空航天大学 | Robust attitude estimation method and application thereof in orientation judgment and related method |
| CN115331263B (en) * | 2022-09-19 | 2023-11-07 | 北京航空航天大学 | Robust attitude estimation method, application of robust attitude estimation method in direction judgment and related method |
| CN115661929A (en) * | 2022-10-28 | 2023-01-31 | 北京此刻启动科技有限公司 | Time sequence feature coding method and device, electronic equipment and storage medium |
| CN115880717A (en) * | 2022-10-28 | 2023-03-31 | 北京此刻启动科技有限公司 | Heatmap key point prediction method and device, electronic equipment and storage medium |
| CN115880717B (en) * | 2022-10-28 | 2023-11-17 | 北京此刻启动科技有限公司 | Heat map key point prediction method and device, electronic equipment and storage medium |
| CN115661929B (en) * | 2022-10-28 | 2023-11-17 | 北京此刻启动科技有限公司 | Time sequence feature coding method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107330439B (en) | Method for determining posture of object in image, client and server | |
| CN109816725B (en) | Monocular camera object pose estimation method and device based on deep learning | |
| CN109063301B (en) | Single image indoor object attitude estimation method based on thermodynamic diagram | |
| Laskar et al. | Camera relocalization by computing pairwise relative poses using convolutional neural network | |
| CN113168510B (en) | Refining shape prior segmentation objects | |
| CN114821263A (en) | Weak texture target pose estimation method based on feature fusion | |
| WO2020108362A1 (en) | Body posture detection method, apparatus and device, and storage medium | |
| Qu et al. | Dilated convolution and feature fusion SSD network for small object detection in remote sensing images | |
| CN111328396A (en) | Pose estimation and model retrieval for objects in images | |
| CN112699806B (en) | Three-dimensional point cloud target detection method and device based on three-dimensional heat map | |
| CN112233181A (en) | 6D pose recognition method and device and computer storage medium | |
| US11704357B2 (en) | Shape-based graphics search | |
| CN111310622A (en) | Fish swarm target identification method for intelligent operation of underwater robot | |
| CN110991513B (en) | Image target recognition system and method with continuous learning ability of human-like | |
| CN113221647B (en) | 6D pose estimation method fusing point cloud local features | |
| CN113159232A (en) | Three-dimensional target classification and segmentation method | |
| CN110781744A (en) | Small-scale pedestrian detection method based on multi-level feature fusion | |
| CN117274388B (en) | Unsupervised three-dimensional visual positioning method and system based on visual text relation alignment | |
| Wang et al. | Revisiting monocular satellite pose estimation with transformer | |
| CN110633727A (en) | Deep neural network ship target fine-grained identification method based on selective search | |
| CN116152334A (en) | Image processing method and related equipment | |
| CN107798329B (en) | CNN-based adaptive particle filter target tracking method | |
| CN110717405A (en) | Face feature point positioning method, device, medium and electronic equipment | |
| CN113436251A (en) | Pose estimation system and method based on improved YOLO6D algorithm | |
| CN117689887A (en) | Workpiece grabbing method, device, equipment and storage medium based on point cloud segmentation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |