373 results about "Block detection" patented technology

Apparatus and methods, such as those that read data from non-volatile integrated circuit memory devices, such as NAND flash. For example, disclosed techniques can be embodied in a device driver of an operating system. Errors are tracked during read operations. If sufficient errors are observed during read operations, the block is then retired when it is requested to be erased or a page of the block is to be written. One embodiment is a technique to recover data from uncorrectable errors. For example, a read mode can be changed to a more reliable read mode to attempt to recover data. One embodiment further returns data from the memory device regardless of whether the data was correctable by decoding of error correction code data or not.

The present invention provides a flash memory management system and method with increased performance. The flash memory management system provides the ability to efficiently manage and allocate flash memory use in a way that improves reliability and longevity, while maintaining good performance levels. The flash memory management system includes a free block mechanism, a disk maintenance mechanism, and a bad block detection mechanism. The free block mechanism provides efficient sorting of free blocks to facilitate selecting low use blocks for writing. The disk maintenance mechanism provides for the ability to efficiently clean flash memory blocks during processor idle times. The bad block detection mechanism provides the ability to better detect when a block of flash memory is likely to go bad. The flash status mechanism stores information in fast access memory that describes the content and status of the data in the flash disk. The new bank detection mechanism provides the ability to automatically detect when new banks of flash memory are added to the system. Together, these mechanisms provide a flash memory management system that can improve the operational efficiency of systems that utilize flash memory.

In certain embodiments this invention pertains to methods of detecting and/or quantifying rare mutant nucleic acids in populations of nucleic acids in which the wild-type nucleic acids are in substantially greater abundance than the rare mutants. In various embodiments the methods utilize short high affinity oligonucleotides targeted to the wild type rather than the minority or mutant sequence. Rather than directly detecting mutant DNA, these probes block detection of wild type DNA. These “blocker” probes can be used in combination with longer “detection” probes or PCR primers to amplify and/or identify the minority mutation in, e.g., clinical specimens. The combination of short high affinity blocker probes and longer, lower affinity detection probes eliminates the single base specificity/complexity tradeoff in the design of nucleic acid probes.

The invention provides a bill automatic identification and processing system based on OCR, which includes an image collection module, a quick image binary module, a text block detection and positioning module, a single column text block accurate positioning module, a multi-column text block accurate positioning and segmentation module, a text identification module, and a bill image retrieval module. The bill automatic identification and processing system based on the OCR can finish the high definition collection and compression storage of invoice bill images, and performs robust and high accurate positioning and identification on literal characters such as information of buying-and-selling sides, goods information, and billing date in each bill. The bill image identification can be conveniently retrieved. The system has the characteristics of high processing and identification precision, low cost, good robustness, and high automation degree, and can be widely used in the computer automatic management of the bills such as the bill authentication and the bill filing and query.

The invention discloses a two-layer block control method of wireless network, comprising that A, current network node detects self block condition and executes step B when in block, B, current network blocked node sends a block control message with self block condition to connected an ascending network node or a nearby node, C, the network node receiving the block control message, according to the block condition adjusts local node transmission speed. The invention makes clear block detection, message sending, and block control, to be directly used in real network, as wireless multi-jump network which comprises wireless network, mobile Ad hoc network and wireless sensor network or the like, thereby reducing local block.

An information processing apparatus capable of accurately detecting commercials included in a television broadcast. A commercial candidate section detector detects a commercial candidate section based on the characteristics of transitions between a program and a commercial and based on the characteristics of commercials. Transitional criteria include quiet sections, scene change points and changes in the audio multiplex mode. A commercial block detector detects a commercial block, which is composed of a plurality of commercial sections and is interposed between program portions, in conformity with the number of frames of the individual commercial candidate sections detected by the commercial candidate section detector, and then outputs to a switch a control signal based on the result of such commercial block detection.

The invention provides a target tracking algorithm based on the self-adaptive particle filter and sparse representation. According to the target tracking algorithm based on the self-adaptive particle filter and the sparse representation, the improved self-adaptive particle filter technique is adopted to serve as a tracking algorithm framework, a block sparse representation model is used for establishing an observation similarity model of a target, partitioning of the target is achieved by means of the self-adaptive partitioning technique, a structural sparse column diagram of a current target state is constructed to calculate the observation similarity of the current target state, blocking is detected by means of a blocking detection mechanism, a target / background dictionary template and a target template column diagram are updated to capture the change of the appearance of the target and the change of the environment during tracking, L1 optimization in the sparse representation is achieved by means of the variable-direction multiplicator method, and then the execution speed of the target tracking algorithm is increased. The target tracking algorithm based on the self-adaptive particle filter and the sparse representation has the advantage that the robustness to the conditions of the posture change of the tracking target, the change of the environment and lighting and blocking is strong.

The invention discloses an anti-blocking target tracking method comprising the following steps: first, the area of a target is determined according to initial image frames, and a tracker forms an initial target template according to the target area; second, the parameters of a cascaded detector are initialized; third, a blocking detection mechanism is added, and the threshold is updated in real time; fourth, the tracking confidence and detection confidence of the tracker and the detector to the target are calculated respectively; and finally, the tracking result is integrated according to the confidence, the tracker is initialized by use of the detection result if the tracker fails to track the target, and the tracking result is input to a learning module in order to update the corresponding parameters of the detector. By adding the blocking detection mechanism, the drift inhibition ability of the tracking algorithm is improved, and lasting and sound tracking is realized.

The invention discloses an abnormal behavior detection method in a video based on target positioning and characteristic fusion. The method comprises motion area detection based on background difference and optical flow statistics, characteristic fusion based on a space-time characteristic and a depth characteristic, and decision based on multi-SVM model training. In the invention, for each dimension characteristic, a classifier is trained respectively, and finally an integration learning method is selected to carry out abnormal detection decision. The invention aims at detecting a motion target area through a rapid detection means, traditional traversing small block detection is improved, calculating efficiency is improved, and simultaneously through fusion of a characteristic layer surface and a model decision layer surface, detection accuracy is increased.

The invention discloses a flash memory and a bad block managing method of the flash memory, and relates to the technical field of NANDFlash memories. The flash memory comprises a bad block detection module, a bad block address table memory module and a block address mapping module. The bad block detection module detects a block address corresponding to access operation initiated by a user and judges whether a block is a bad block or not, the bad block address table memory module conducts memory on the block address of the bad block judged by the bad block detection module, when the block address mapping module receives the access operation initiated by the user, the block address mapping module inquires whether the block address, corresponding to the access operation, in the logic address is the block address of the bad block or not from the bad block address table memory module, and if the block address, corresponding to the access operation, in the logic address is the block address of the bad block, the block address, corresponding to the access operation, in the logic address is mapped to the block address of a good block, and then is sent into the flash memory to conduct address decoding. The invention further discloses a bad block managing method by the flash memory. According to the technical scheme, the reliability of the system is improved, and the bottleneck of a SPI NAND Flash application is overcome.

The invention relates to a firmware upgrade method and a firmware upgrade device for an embedded system. The method comprises the following steps: (a) acquiring and storing an upgrade mirror file to a first partition area of a nonvolatile memory of the embedded system, and generating the storage path information of the upgrade mirror file; (b) performing bad block detection on a second partition area of the nonvolatile memory; (c) copying the upgrade mirror file from the first partition area to the good block of the second partition according to the storage path information and modifying the start position of the operating system to be the position of the second partition; (d) performing bad block detection on a third partition of the nonvolatile memory after the operating system is restarted from the second partition in the embedded system; (e) copying the upgrade mirror file to a good block of the third partition area, and modifying the start position of the operating system to be position of the third partition area; and (f) restarting the operating system from the third partition area in the embedded system. Therefore, the upgrade failure problem brought by the bad block generated in the nonvolatile memory such as NAND FLASH in the operating system upgrading process can be solved.

The invention discloses a floor defect automatic detection method based on computer vision and deep learning, which comprises the following steps of: before detection, making a sample by utilizing anexisting floor defect image, and training a floor defect detection model according to a transfer learning theory; the detection process comprises the following steps: firstly, segmenting and extracting a floor image wall body area according to traditional image processing; secondly, partitioning a wall body area of the floor image, detecting each sub-block by applying a model, and accelerating byadopting a parallel processing mode; integrating and converting a sub-block detection result, and marking the position of the defect on the original image; and finally, correcting the detection result, and making a new sample for re-training the model to improve the detection performance of the model. The problems that in the prior art, the speed and precision of floor image defect detection are difficult to achieve at the same time, the adaptability of a detection algorithm is not high, and the detection effect is easily affected by the photo shooting environment are solved.

The invention provides a bad block management method for a flash memory. The method comprises the following steps of: dividing the storage space of the flash memory into a first area comprising a plurality of blocks and a second area comprising a plurality of blocks; performing bad block detection on the flash memory to acquire information about bad blocks; establishing a mapping relation between the bad blocks in the first area and normal blocks in the second area according to the acquired information about the bad blocks, and generating a block mapping quick access table comprising the information about the established mapping relation; and respectively storing the generated block mapping quick access table into a plurality of bad block management information blocks in the second area, wherein a plurality of bad block management information blocks are any normal blocks which are positioned in the second area and do not have the mapping relation with the bad blocks in the first area.

A storage system according to the present invention includes: a block detecting means for checking whether respective block data obtained by division are already stored in a storage device; and a data writing means for storing the respective block data obtained by division with duplicated storage eliminated into the storage device. The block detecting means detects a share rate representing a rate of a shared portion between a plurality of consecutive block data configuring a predetermined range in write target data among the block data obtained by division and a plurality of block data in a predetermined range already stored consecutively in the storage device. The data writing means, depending on the detected share rate, newly stores the block data obtained by division, into the storage device.

The invention discloses a road blocking information extraction based on deep learning image classification, and the method comprises the steps: building a road blocking image classification sample library through employing a disaster typical case image, carrying out the training of a convolutional neural network, and obtaining an initial convolutional neural network model CNNmodel0; obtaining a post-disaster image I (x) and a road vector R (x) of the research area x, and detecting road blocking to obtain a to-be-detected sample Dn (x); using the trained network CNNmodel0 to perform multi-pointroad blocking detection of adding detection points in the vertical normal direction of the road vector direction on the sample Dn (x) to be detected: selecting a small number of road segments to generate new training sample data S1, and performing network fine tuning on the existing network; and repeating the road blocking classification detection work until the detection result meets the precision requirement. Various precision evaluation indexes of the improved convolutional neural network model are superior to those of an original model, and the improved convolutional neural network modelis more suitable for specific problems of disaster area road blocking image classification and detection.

The invention provides a method for detecting all-zero blocks in H.264. Based on the relativity between a reference block and a current encoding block and by combining with an idea of mode recognition, a linear classifier for all-zero block detection is provided. Through carrying out all-zero block prejudgment before transform and quantization, the condition that a residual block coefficient is zero can be detected in advance, thereby omitting the transform and quantization process of a residual block. The method has an extremely great significance for encoding of a video with a low code rate, because the proportion of the all-zero blocks is very large under the condition of low code rate, and the calculation complexity of the transform and quantization can be greatly reduced if most of all-zero blocks are detected, thereby improving encoding efficiency.

A device for giving importance information according to a video operation history includes an interest video searching operation block detection unit provided with a minimum searching operation block detection unit for detecting a minimum searching operation block as a minimum unit which exists in an interest video searching operation block from a preview log as a video operation history of a user, and a searching operation block setting unit for setting the interest video searching operation block through determination with respect to an operation previous and subsequent to the detected minimum searching operation block as a base point based on a feature of the interest video searching operation block.

The invention relates to the field of computer vision recognition, and provides an image description method constructed based on a hierarchical feature relation graph, which comprises the following steps: constructing a training data set; inputting the image into a block detection module, and outputting block visual information; inputting the image into a target detection module, and outputting target visual information; inputting the image into a text detection module, and outputting text visual information; Respectively inputting the three visual information into a description generator, respectively constructing a relational graph of each type of visual information and training image coordinate information, and optimizing the three visual information; Screening and fusing the three types of visual information to obtain multi-modal characteristics; inputting the description words into a recurrent neural network to extract feature information, and predicting the next description wordto generate a complete description sentence. According to the method, various types of visual information are optimized, screened and fused, so that any input test image is described, and the image description accuracy can be effectively improved.

The invention discloses a NAND Flash fault-tolerance method based on a FPGA (Field Programmable Gate Array). The NAND Flash fault-tolerance method comprises the following steps: carrying out Flash bad block detection, bad block management and dynamic table lookup. The bad block is detected through judging whether the state of a Flash write zone bit is the same with read-write contents or not. The addresses of detected good blocks and bad blocks are managed by a bad block management algorithm, and an address mapping table is established. The FPGA carries out write, read and erasure operations on the Flash through a dynamic table lookup way, and the accuracy and the efficiency of data access are greatly improved. The method can directly improve the space use ratio of the Flash, the integral performance and the processing speed of the Flash can obtain an important guarantee, and the method has a reference meaning for the application development of the new-generation large-volume NAND Flash.

The embodiment of the invention discloses a prompt method and device, a storage medium, and a mobile terminal. The method comprises the following steps: when a blocking detection event is triggered, acquiring a photographing preview image; inputting the photographing preview image into a pre-trained blocking detection model; determining whether a first blocked area exists in the photographing preview image or not according to an output result of the blocking detection model; and prompting a user to remove a blocking object when it is determined that the first blocked area exists in the photographing preview image, wherein the blocking object comprises an object that causes the existence of the first blocked area in the photographing preview image. The prompt method and device, the storagemedium and the mobile terminal adopting the technical scheme have the advantages that blocking detection is carried out on the photographing preview image through the pre-constructed blocking detection model, whether the blocked area exists in the photographing preview image or not is judged accurately and rapidly, and the user is prompted in time to remove the blocking object when it is determined that the first blocked area exists in the photographing preview image, so that the quality of a photographed image can be effectively improved.

Methods and apparatus implementing systems and techniques for differencing token streams and detecting moved blocks of tokens. In general, in one implementation, the technique includes: obtaining a first token stream and a second token stream, comparing the first and second token streams to identify a group of tokens that are substantially similar in the first and second token streams, the similar-tokens group including common sub-sequences, which are identical in the first and second token streams, and at least one unmatched token, and presenting matched token information corresponding to the similar-tokens group to represent changes in document flow.

The invention discloses a method and device for detecting blocking rotation of a direct current brush motor, wherein the method comprises the following steps: in each starting stage of the direct current brush motor, the blocking rotation detection threshold value is obtained according to the current starting characteristic analysis of the direct current brush motor, and the detection is carried out according to the blocking rotation detection threshold value in the subsequent blocking rotation detection. As the current starting characteristic of the direct current brush motor is analyzed during each starting process of the direct current brush motor to obtain the blocking detection threshold value, the blocking detection threshold value can be suitable for the direct current brush motor under different temperatures and voltages, thereby improving the practicability and the reliability of the technical scheme, and reducing the calibration workload due to no need of calibrating each direct current brush motor.

The invention discloses a blocking detection method of an external circulation air inlet of a heat exchanger and a system thereof. The method comprises: a heat exchanger detects the rotation speed ofan external circulation fan and sends a detected rotation speed value to a network management center; the network management center compares the rotation speed value with a preset range so as to judgewhether the external circulation air inlet is blocked or not. The invention can solve the problems that the heat exchanger needs dedusting periodically on a base station site in the relevant technology and maintenance cost is high; thus, the invention can accurately and remotely judge the state of the external circulation air inlet of the heat exchanger and avoid aimless maintenance of the heat exchanger on the base station site, so that the invention not only effectively cleans the heat exchanger but also lowers the maintenance cost of equipments.

The invention discloses a cloud storage method for reducing a data-occupied space. The method comprises the following steps: uploading data to a cloud storage server from a terminal; using a block detection technique of a fixed size classification algorithm to delete the repeated data; and after the data is stored into a data block, using an LZW data compression technique to reduce the size of a basic data block so as to realize data compression. The data compression technique and the repeated data deleting technique are combined to delete the repeated data and then compress the data, the higher data reduction ratio is realized, and the higher data compression ratio and data compression performance are obtained.