International Business Machines Corporation

Guoqiang Hu

Jian Ding

Ning Duan

Bo Yang

Peng Ji

Li Jun Mei

Jian Wang

Qi Cheng Li

Yu Fen Yuan

Yi Min Wang

Junchi Yan

Jian Dong Ding

HuaFang Tan

Zhi Hu Wang

Wei Shan Dong

Peng Gao

Yi Qi

Min Gong

Yu Ying Wang

Sheng Huang

Yuan Yuan Ding

Zi Yu Zhu

Sheng Nan Zhu

Xin Hao Cheng

JingChang Huang

Jiandong Ding

Xin Zhang

Charles Daniel Wolfson

Fan Jing Meng

Dan Zhang

Jing Chang Huang

Fan Li

Jian Xu

Hong Bing Zhang

Ting Yuan

Tian Tian Chai

Philip L Schwartz

Chen Hao

A method for detecting and labeling a target object in a 2D image includes receiving a plurality of 2D images from a visual sensor, manually marking points of the target object on each of the 2D images, generating from the 2D images a 3D world coordinate system of the environment surrounding the target object, mapping each of the marked points on the 2D images to the 3D world coordinate system using a simultaneous localization and mapping (SLAM) engine, automatically generating a 3D bounding box covering all the marked points mapped to the 3D world coordinate system, mapping the 3D bounding box to each of the 2D images, generating a label for the target object on each of the 2D images using a machine learning object detection model, and training the machine learning object detection model based on the generated label for the target object.

AR-based labeling tool for 3D object detection model training

A system, method, and computer program product provides a way to separate connected or adhered adjacent characters of a digital image for license plate recognition. As a threshold processing, the method performs a recognition of character adhesion by obtaining character parameters using an image processor. The parameters include a horizontal max crossing and a ratio of width and height. A first rule-based module is used responsive to the character parameters to distinguish the adhered characters (character adhesions) that are easy to judge, leaving the uncertain part to a character adhesion classifier model for discrimination. Character adhesion data is obtained by data augmentation including the adding of a random distance between two single characters to create class like adhered characters. Then the character adhesion classifier model of single character and character adhesion data is trained. Any uncertain part can be distinguished by the trained character adhesion classifier model.

Character recognition of license plate under complex background

A method, system, and computer program product for obtaining a first route traversed by a target object, performing at least one prediction for a second route to be traversed by the target object based on the first route, the at least one prediction being performed with at least one of an object-specific prediction model, an object group-specific prediction model, and an object-independent prediction model, and determining, according to a decision rule, a prediction result of the second route based on the at least one prediction.

Route prediction based on adaptive hybrid model

A method and an apparatus for generating a road network are disclosed. The method for generating a road network comprises: aggregating a plurality of grid cells partitioned in advance on a trajectory map based on trajectories in each grid cell of the plurality of grid cells to form level-1 regions; and generating a link of the road network by merging a level-1 region having two valid neighbors with its neighbor level-1 regions having two valid neighbors.

Road network generation

One or more sensors associated with a vehicle detect a roadway condition of a first roadway and an evasive maneuver is performed by the vehicle to avoid the detected roadway condition. In response to one or more processors determining that the evasive maneuver was successful, a record of the successful maneuver and the roadway condition are stored in a database. Subsequent to storing the record in the database, one or more computers associated with one or more vehicles is trained to execute the evasive maneuver, in response to determining that the one or more vehicles is exposed to the roadway condition experienced by the vehicle.

Training a self-driving vehicle

An apparatus or a method for detecting activity quantity of an object includes a sense unit, a count unit, a trigger unit, and a radio circuit. The sense unit senses movements of the object, and generates a sensing signal based on the sensed movements. The count unit receives the sensing signal from the sense unit, counts a number of the movements from the sensing signal, and generates a movement number signal including the number of the movements. The trigger unit generates a trigger signal based on whether one or more trigger conditions are satisfied. The radio circuit is triggered in response to the trigger signal, to generate an activity quantity signal based on the movement number signal and ID information corresponding to the object.

Sensor based activity monitor

Managing software performance debugging based on a distributed VM system is provided. In response to determining a debugging state of a software system running on a VM, a timing of a system clock of the VM is controlled. A data packet sent to the VM from another VM is intercepted, and an added system time and reference time that indicate when the packet was sent by the other VM is extracted from the packet. Based on the system and reference times, as well as a reference time of when the packet is intercepted, a timing at which the packet is expected to be received by the VM is calculated. The packet is forwarded to the VM as a function of a comparison of the timing at which the packet is expected to be received and a system time of the VM when the packet is intercepted.

Managing software performance tests based on a distributed virtual machine system

Embodiments of the present invention provide information processing systems and methods for proliferating testing data based on sample testing data. In one embodiment, a difference is determined by comparing a desired query result to a query result acquired by executing a query statement on sample testing data. Sample testing data can then be proliferated based, at least in part, on the difference and data generation constraint conditions.

Method and apparatus for proliferating testing data

Embodiments of the present invention disclose a technical solution of determining a border between road network partitions, comprising: determining a border point based on road network data and partitioning information of road segments in the road network; determining a buffer zone between adjacent partitions based on the border point, the partitioning information of the road segments, and the road network data; determining a border between the adjacent partitions based on the buffer zone. With the solution of the present invention, border lines for road network partitions can be accurately determined.

Method and apparatus for determining a road network partitioning border line

Methods and systems for monitoring interesting subjects. A method including: selecting, based on a first collection of interesting subjects, a set of critical nodes including at least one critical node which participates in one or more interesting subjects in the first collection; and monitoring contents posted by the one or more critical nodes in the set so as to find a second collection of interesting subjects. The set of critical nodes which participate in one or more interesting subjects in the first collection of interesting subjects is selected based on the first collection, as objects to be monitored, thereby reducing the number of contents posted by the nodes to be monitored as compared with monitoring all the user nodes, so that interesting subjects such as hot news or hot events can be found in real time with high efficiency and low cost.

Monitoring interesting subjects

Simplification of trajectory representation is provided. Trajectories represented by a set of points on a map are obtained. Common route elements are identified for the trajectories based on the set of points. Each of the common route elements represents a geospatial area in which at least a predetermined number of the trajectories are identical. Simplified representations of the trajectories are generated based on the identified common route elements.

Simplification of trajectory representation

A method for detecting the presence of an unattended child left in a vehicle includes receiving acoustic data obtained from inside a vehicle, determining if the vehicle is running or not in response to the acoustic data, determining if all windows in the vehicle are closed when the vehicle is not running, determining if sound volume inside the vehicle exceeds a predetermined threshold when all of the windows are closed, determining if a predetermined high-frequency word is spoken inside the vehicle when all of the windows are closed, and outputting an alert indicating that an unattended child is left in the vehicle when the sound volume inside the vehicle exceeds the predetermined threshold and when the predetermined high-frequency word is spoken inside the vehicle.

Method, device and system for detecting the presence of an unattended child left in a vehicle

Dispatching messages in a moving object, in one aspect, may include receiving a description of a spatial area. A hierarchical tree representation of the spatial area may be generated. The hierarchical tree representation may comprise nodes, wherein a child node represents a sub-region of a region represented by a parent node of the child node. Moving features may be associated with each of the nodes, the moving features comprising whether the node is cacheable, whether the node covers an unpartitioned area, an active object list in an area covered by the node, and a latest position associated with an object in the active object list. Data from the moving object may be dispatched based on the hierarchical tree representation.

High performance and scalable telematics message dispatching

Growing community of inventors

Shanghai, China

Guoqiang Hu