A multimodal AI method and recognition system for job assessment in the power industry

Abstract

This invention relates to a multimodal AI method and recognition system for job assessment in the power industry, belonging to the field of digital data processing technology. The method includes the following steps: Step 1): Training a SAM model and an ASR model that meet the requirements of the test scenario; Step 2): In the SAM model, constructing a visual component that combines target recognition atomic components, state recognition atomic components, and position change atomic components; Step 3): In the ASR model, constructing a speech component; Step 4): Combining the visual and speech components to achieve intelligent job assessment and evaluation. The beneficial effects of this invention are that the images or videos recognized by the SAM model and the language and text recognized by the ASR model are assessed and evaluated through the assessment module, improving the efficiency and quality of intelligent and digital job assessment and evaluation.

Description

Technical Field

[0001] This invention belongs to the field of electrical digital data processing technology, and specifically relates to a multimodal AI method and recognition system for job evaluation in the power industry. Background Technology

[0002] The assessment and evaluation of professional skills and job competency plays a crucial role in power industry enterprises. It is not only a primary benchmark for measuring employee training effectiveness and skill levels, but also an indispensable core basis for the scientific selection and appointment of talent within power industry enterprises. This system not only concerns the career development of individual employees but also directly relates to the overall operational efficiency and service quality of power industry enterprises.

[0003] However, faced with the continuous emergence of new job positions and business models, as well as the rapid changes in the market environment, the existing job evaluation system still faces many challenges and urgently needs to be optimized and upgraded in the following key aspects:

[0004] (1) The traditional on-site job assessment model has an efficiency bottleneck;

[0005] (2) The organization and coordination of traditional job evaluation work is cumbersome and complicated;

[0006] (3) The direct interaction between candidates and examiners in the traditional model carries interaction risks;

[0007] (4) The organization of job evaluation activities has a significant impact on grassroots work.

[0008] With the successful application of big data, artificial intelligence, and other information and digital technologies across various industries, the power industry has also ushered in new opportunities for digital transformation. Introducing advanced technologies can not only effectively improve the efficiency and quality of job evaluation, but also reduce human and material costs and enhance the objectivity and fairness of evaluations. Therefore, power industry enterprises should actively explore the path of digital transformation, integrating information and digital technologies into job evaluation work, driving development through innovation, continuously improving human resource management, and providing strong support for the sustainable development of the enterprise. Summary of the Invention

[0009] This invention provides a multimodal AI method and recognition system for job assessment in the power industry. It addresses the technical problem of how to digitally improve the efficiency and quality of job assessment and evaluation in existing technologies. By combining visual and voice components, the system performs in-depth data analysis. Specifically, the system deeply mines and analyzes the candidate's operational data and the candidate's voice, thereby improving the efficiency and quality of intelligent and digital job assessment and evaluation.

[0010] To achieve the above objectives, the present invention is implemented through the following technical solution:

[0011] A multimodal AI method for job assessment in the power industry includes the following steps:

[0012] Step 1): Train the SAM model and ASR model to meet the requirements of the test question scenario; the SAM model is used for image or video recognition, and the ASR model is used for recognizing the speech expressed by the test taker into written language.

[0013] Step 2): In the SAM model, construct a visual component that combines target recognition atomic components, state recognition atomic components, and position change atomic components;

[0014] Step 3): In the ASR model, construct the speech-related components;

[0015] Step 4): Combine visual and voice components to achieve intelligent job assessment and evaluation.

[0016] Optionally, in step 2), the scores of visual components are recorded in the test questions, wherein the scores of visual components in the test questions are recorded as follows: .

[0017] Optionally, in step 2), the target recognition atomic component identifies the spatial relationship between two targets in an image or video, specifically through the following steps:

[0018] Step 01): Set two spatial position relationship thresholds for each of the two targets. and ;

[0019] Step 02): Calculate the predicted bounding box for a target. And the prediction box of another target Crossover ratio between The intersection-union ratio is calculated using formula (1): (1);

[0020] Step 03): Calculate the depth value between the two targets. The depth value between two targets is calculated using formula (2): (2); In formula (2), The depth value that the SAM model can predict. It is the maximum depth value among the depth values ​​between two targets, used for normalization; It is a scaling factor used to normalize the maximum depth value. Restored to the original scale;

[0021] Step 04): Based on the intersection-union ratio The threshold for the spatial positional relationship between each target and its respective target , Comparison based on depth values ​​between two targets Thresholds for the spatial positional relationship between each target and its respective target , The comparison is used to determine the recognition result of the target recognition atomic component. .

[0022] Furthermore, in step 04), the identification result of the target identification atomic component. The following formula (3) is adopted:

[0023] (3);

[0024] In formula (3), if and The predicted bounding box of the target And the prediction box of the other target The crossover ratio Greater than the spatial position relationship threshold The depth value between the two targets Less than the spatial position relationship threshold If the recognition is successful, then the recognition is successful. If the value is 1, otherwise the recognition fails and It is 0.

[0025] Optionally, in step 2), the state recognition type atomic component recognition involves identifying two labels for a target at different time points. The specific steps are as follows:

[0026] Step a): Implement two states for a target. and The two types of labels were identified respectively, and the identification result sequences for the two types of labels were as follows: and ;

[0027] Step b): In Time's up At any given time, the recognition result of either 0 or 1 for the two labels is appended to the... The result sequence and the above In the result sequence;

[0028] Step c): In At any time, for Result sequence and The two sequences in the result sequence are XORed, and the result is then converted into decimal data for comparison; wherein, formula (4) is used to convert the result of the XOR into decimal data for comparison: (4); In formula (4), For the XOR operation, The operation of converting the result of the XOR operation into decimal data. The result of the XOR operation is then converted to decimal data for comparison; if A value of 1 indicates successful recognition; if A value of 0 indicates recognition failure.

[0029] Optionally, in step 2), the identification of position-change type atomic components is used to determine the spatiotemporal relationship of a target at different times. The specific steps are as follows:

[0030] Step A): First, set a target's position change distance threshold. and position change angle threshold ;

[0031] Step B): In Time's up At any given moment, locate the position of a target within the field of view in space, where the positions within the field of view are respectively... Coordinates of the predicted bounding box of an object at a given time. ,as well as Coordinates of the predicted bounding box of an object at a given time. ;

[0032] Step C): Calculate the Euclidean distance between the two coordinate points. Specifically, it is as follows: (5)

[0033] (5);

[0034] Step D): Select The axis is horizontal; calculate the angle between two coordinate points. Specifically, it is the following formula (6):

[0035] (6);

[0036] Step E): Calculate the recognition result of the spatiotemporal relationship of a target at different times. The formula is as follows (7):

[0037] (7);

[0038] in, and ,but A value of 1 indicates successful recognition; otherwise, If the value is 0, recognition failed.

[0039] Optionally, in step 3), the score for the speech component is recorded in the test questions, wherein the score for the speech component in the test questions is recorded as follows: .

[0040] Optionally, in step 4), the score for the test question is calculated using the following steps:

[0041] Step I): The score of the visual component is calculated as the confidence level of the image or video segment recognized by the SAM model. Multiply by the allocation of visual components ;

[0042] Step II): The score of the speech component is calculated as the sum of the weights of the keywords in the language text recognized by the ASR model. Multiply by the allocation of voice components ;in, For the identified first One keyword, The total number of keywords identified;

[0043] Step III): Final score of the exam questions The calculation method is as follows, as shown in formula (8):

[0044] (8);

[0045] in, For the first Visual-related test points To identify the first Confidence level of images or video clips for each visual test point , For the first The allocation of visual components for each visual test point , For a test question there are Visual-related test points for The score of visual components for each visual test point;

[0046] For the first One key point related to pronunciation. For the first The allocation of points for speech-related components in each speech-related test point , For a test question there are One key point related to pronunciation. for The score of the voice component for each voice-related test point.

[0047] A recognition system based on a multimodal AI method for job assessment in the power industry, comprising:

[0048] SAM model, used for image or video recognition;

[0049] The ASR model is used to recognize and convert the speech expressed by test takers into written language.

[0050] The performance evaluation module is used for job performance evaluation.

[0051] The assessment and evaluation module is connected to the SAM model and the ASR model. The image or video recognition data in the SAM model and the language and text recognition data in the ASR model are sent to the assessment and evaluation module, which then assesses and evaluates the job.

[0052] Optionally, the SAM model includes: a target recognition atomic component module for identifying the spatial relationship between two targets in an image or video; a recognition atomic component module for identifying two label states of a target at different time points; and a position change atomic component module for identifying and judging the spatiotemporal relationship of a target at different times.

[0053] The ASR model includes: speech-related component modules for speech-to-text conversion and recognition;

[0054] The assessment and evaluation module is connected to the target recognition atomic component module, the atomic component module, and the position change atomic component module; the assessment and evaluation module is connected to the voice component module.

[0055] The assessment and evaluation module combines visual components (a combination of target recognition atomic components, atomic components, and position change atomic components) with voice components from the voice component module for job assessment and evaluation.

[0056] The beneficial effects of this invention are:

[0057] This invention utilizes visual components to digitally record operational details and voice components to record voice segments. The visual and voice components are combined for in-depth data analysis: the system performs in-depth mining and analysis of the examinee's operational data and voice data, thereby improving the efficiency and quality of intelligent and digital job assessment and evaluation. Attached Figure Description

[0058] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0059] Figure 1 This is a schematic diagram of the identification system structure of the present invention;

[0060] Figure 2 This is a flowchart of the process of the present invention;

[0061] Figure 3 This is a schematic diagram illustrating the recognition method of the visual components of the present invention;

[0062] Figure 4 This is a schematic diagram illustrating the recognition method of the voice component of the present invention. Detailed Implementation

[0063] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0064] Example 1;

[0065] like Figure 1 As shown, this embodiment provides a recognition system for a multimodal AI method applied to job assessment in the power industry, including:

[0066] SAM model, used for image or video recognition;

[0067] The ASR model is used to recognize and convert the speech expressed by test takers into written language.

[0068] The performance evaluation module is used for job performance evaluation.

[0069] The assessment and evaluation module is connected to the SAM model and the ASR model. The image or video recognition data in the SAM model and the language and text recognition data in the ASR model are sent to the assessment and evaluation module, which then assesses and evaluates the job.

[0070] The SAM model includes: a target recognition atomic component module for identifying the spatial relationship between two targets in an image or video; a recognition atomic component module for identifying two label states of a target at different time points; and a position change atomic component module for identifying and judging the spatiotemporal relationship of a target at different times.

[0071] The ASR model includes: speech-related component modules for speech-to-text conversion and recognition;

[0072] The assessment and evaluation module is connected to the target recognition atomic component module, the atomic component module, and the position change atomic component module; the assessment and evaluation module is connected to the voice component module.

[0073] The assessment and evaluation module combines visual components (a combination of target recognition atomic components, atomic components, and position change atomic components) with voice components from the voice component module for job assessment and evaluation.

[0074] In this embodiment, the images or videos recognized by the SAM model and the language and text recognized by the ASR model are evaluated through an assessment module.

[0075] Example 2;

[0076] Based on Example 1, such as Figure 2 As shown, this embodiment provides a multimodal AI method for job assessment in the power industry, including the following steps:

[0077] Step 1): Train the SAM model and ASR model to meet the requirements of the test question scenario; the SAM model is used for image or video recognition, and the ASR model is used for recognizing the speech expressed by the test taker into written language.

[0078] Step 2): In the SAM model, construct a visual component that combines target recognition atomic components, state recognition atomic components, and position change atomic components;

[0079] Step 3): In the ASR model, construct the speech-related components;

[0080] Step 4): Combine visual and voice components to achieve intelligent job assessment and evaluation.

[0081] In step 2), the scores of visual components are recorded in the exam questions, where the scores of visual components in the exam questions are recorded as follows: .

[0082] In step 2), the target recognition atomic component identifies the spatial relationship between two targets in an image or video (i.e., the spatial relationship between objects, or the spatial relationship between a human joint (hand, foot, head, etc.) and an object), specifically through the following steps:

[0083] Step 01): Set two spatial positional relationship thresholds for each of the two targets. and ;

[0084] Step 02): Calculate the predicted bounding box for a target. And the prediction box of another target Crossover ratio between The intersection-union ratio is calculated using formula (1): (1);

[0085] Step 03): Calculate the depth value between the two targets. The depth value between two targets is calculated using formula (2): (2); In formula (2), The depth value that the SAM model can predict. It is the maximum depth value among the depth values ​​between two targets, used for normalization; It is a scaling factor used to normalize the maximum depth value. Restored to the original scale;

[0086] Step 04): Based on the intersection-union ratio Thresholds for the spatial positional relationship between each target and its respective target , Comparison based on depth values ​​between two targets Thresholds for the spatial positional relationship between each target and its respective target , The comparison is used to determine the recognition result of the target recognition atomic component. .

[0087] Furthermore, in step 04), the identification result of the target identification atomic component... The following formula (3) is adopted:

[0088] (3);

[0089] In formula (3), if and A prediction box for a target And the prediction box of another target intersection ratio Greater than the spatial relationship threshold Depth value between two targets Less than the spatial relationship threshold If the recognition is successful, then the recognition is successful. If the value is 1, otherwise the recognition fails and It is 0.

[0090] In step 2), the state recognition atomic component identification involves identifying two labels for a target at different time points. The specific steps are as follows:

[0091] Step a): Implement two states for a target. and (Because a target has different states at different times) Two types of labels were identified, and the identification result sequences for the two types of labels are as follows: and ;

[0092] Step b): In Time's up At any given moment, for both labels, the recognition result of 0 or 1 is appended to... The result sequence and In the result sequence;

[0093] Step c): In At any time, for Result sequence and The two sequences in the result sequence are XORed, and the result is then converted into decimal data for comparison; wherein, formula (4) is used to convert the result of the XOR into decimal data for comparison: (4); In formula (4), For the XOR operation, The operation of converting the result of the XOR operation into decimal data. The result of the XOR operation is then converted to decimal data for comparison; if A value of 1 indicates successful recognition; if A value of 0 indicates recognition failure.

[0094] In step 2), the identification of position-change atomic components determines the spatiotemporal relationship of a target at different times. The specific steps are as follows:

[0095] Step A): First, set a target's position change distance threshold. and position change angle threshold ;

[0096] Step B): In Time's up At any given moment, locate the position of a target within the field of view in space. The positions within the field of view are as follows: Coordinates of the predicted bounding box of an object at a given time. ,as well as Coordinates of the predicted bounding box of an object at a given time. ;

[0097] Step C): Calculate the Euclidean distance between the two coordinate points. Specifically, it is as follows: (5)

[0098] (5);

[0099] Step D): Select The axis is horizontal; calculate the angle between two coordinate points. Specifically, it is the following formula (6):

[0100] (6);

[0101] Step E): Calculate the recognition result of the spatiotemporal relationship of a target at different times. The formula is as follows (7):

[0102] (7);

[0103] in, and ,but A value of 1 indicates successful recognition; otherwise, If the value is 0, recognition failed.

[0104] In step 3), the score for the voice component is recorded in the test questions, where the score for the voice component in the test questions is recorded as follows: .

[0105] In step 4), the score for the test questions is calculated using the following steps:

[0106] Step I): The score for visual components is calculated based on the confidence level of the image or video segment recognized by the SAM model. Multiply by the allocation of visual components ;

[0107] Step II): The score for the speech component is calculated as the sum of the weights of the keywords in the language text recognized by the ASR model. Multiply by the allocation of voice components ;in, For the identified first One keyword, The total number of keywords identified;

[0108] Step III): Final score of the exam questions The calculation method is as follows, as shown in formula (8):

[0109] (8);

[0110] in, For the first Visual-related test points To identify the first Confidence level of images or video clips for each visual test point , For the first The allocation of visual components for each visual test point , For a test question there are Visual-related test points for The score of visual components for each visual test point;

[0111] For the first One key point related to pronunciation. For the first The allocation of points for speech-related components in each speech-related test point , For a test question there are One key point related to pronunciation. for The score of the voice component for each voice-related test point.

[0112] Example 3;

[0113] Based on Example 2, such as Figure 3 As shown, target recognition atomic components, state recognition atomic components, and position change atomic components are one or more atomic components combined in an ordered or unordered manner, capable of meeting a specific assessment requirement, and are stateless. Take the vision category (inspecting a safety helmet) as an example.

[0114] 1. The inspection of the safety helmet is broken down into three atomic components (i.e., target recognition atomic component, state recognition atomic component, and position change atomic component): Hold the safety helmet with both hands, with one hand holding the safety helmet and the other hand pointing to the safety helmet certificate; or hold the safety helmet with one hand and squeeze the helmet support with the other hand.

[0115] 2. Drag the three atomic components mentioned above to the arrangement area and set the attributes for the tags and time order.

[0116] 3. Publish the arranged visual components to the component marketplace for use in subsequent exam question arrangement.

[0117] This embodiment is a visual helmet inspection. By recognizing images or videos of both hands or both hands, it determines the score for the examinee's visual components. The SAM model automatically identifies the confidence level of an image or video clip. The scores for visual components are then calculated.

[0118] Example 4;

[0119] Based on Example 2, such as Figure 4 As shown, taking speech-related questions as an example: converting spoken language into written language. For instance, the test point about wearing a safety helmet properly requires assigning weights to the phrase "the safety helmet should not slip down when tilted down, and should not loosen when tilted up," as well as keywords such as "safety helmet," "slip down," and "not loosen."

[0120] This embodiment calculates the score for the candidate's voice-related components by converting speech into written text. The sum of the weights of keywords in the language text automatically identified by the ASR model The score for the voice component is then calculated.

[0121] Example 4;

[0122] Based on all the above embodiments, the present invention is a visual component for digitally recording operational details: the system can accurately capture and identify every operational detail of the examinee, save the operational details in digital form, and form a complete and detailed operation record.

[0123] The voice component records voice segments. The system can accurately capture and recognize the examinee's voice, convert the captured voice into written language, and save the written language in digital form to form a complete and detailed voice record.

[0124] By combining visual and voice components, the system conducts in-depth data analysis: it performs deep mining and analysis of candidates' operational data and voice data, thereby improving the efficiency and quality of intelligent and digital job assessment and evaluation.

[0125] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope described in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A multimodal AI method for job assessment in the power industry, characterized in that, Includes the following steps: Step 1): Train the SAM model and ASR model to meet the requirements of the test question scenario; the SAM model is used for image or video recognition, and the ASR model is used for recognizing the speech expressed by the test taker into written language. Step 2): In the SAM model, construct a visual component that combines target recognition atomic components, state recognition atomic components, and position change atomic components; In step 2), the target recognition atomic component identifies the spatial relationship between two targets in an image or video, specifically through the following steps: Step 01): Set two spatial position relationship thresholds for each of the two targets. and ; Step 02): Calculate the predicted bounding box for a target. And the prediction box of another target Crossover ratio between The intersection-union ratio is calculated using formula (1): (1); Step 03): Calculate the depth value between the two targets. The depth value between two targets is calculated using formula (2): (2); In formula (2), The depth value that the SAM model can predict. It is the maximum depth value among the depth values ​​between two targets, used for normalization; It is a scaling factor used to normalize the maximum depth value. Restored to the original scale; Step 04): Based on the intersection-union ratio The threshold for the spatial positional relationship between each target and its respective target , The comparison is based on the depth values ​​between the two targets. Thresholds for the spatial positional relationship between each target and its respective target , The comparison is used to determine the recognition result of the target recognition atomic component. ; Step 3): In the ASR model, construct the speech-related components; Step 4): Combine visual and voice components to achieve intelligent job assessment and evaluation.

2. The multimodal AI method for job assessment in the power industry according to claim 1, characterized in that, In step 2), the score of the visual component is recorded in the exam question, wherein the score of the visual component in the exam question is recorded as follows: .

3. The multimodal AI method for job assessment in the power industry according to claim 1, characterized in that, In step 04), the recognition result of the target recognition atomic component. The following formula (3) is adopted: (3); In formula (3), if and The predicted bounding box of the target And the prediction box of the other target The crossover ratio Greater than the spatial position relationship threshold The depth value between the two targets Less than the spatial position relationship threshold If the recognition is successful, then the recognition is successful. If the value is 1, otherwise the recognition fails and It is 0.

4. The multimodal AI method for job assessment in the power industry according to claim 1, characterized in that, In step 2), the state recognition atomic component recognition involves identifying two labels for a target at different time points. The specific steps are as follows: Step a): The target is placed in two states. and The two types of labels were identified respectively, and the identification result sequences for the two types of labels were as follows: and ; Step b): In Time's up At any given time, the recognition result of either 0 or 1 for the two labels is appended to the... The result sequence and the above In the result sequence; Step c): In At any time, for Result sequence and The two sequences in the result sequence are XORed, and the result is then converted into decimal data for comparison; wherein, formula (4) is used to convert the result of the XOR into decimal data for comparison: (4); In formula (4), For the XOR operation, The operation of converting the result of the XOR operation into decimal data. The result of the XOR operation is then converted to decimal data for comparison; if A value of 1 indicates successful recognition; otherwise... A value of 0 indicates recognition failure.

5. The multimodal AI method for job assessment in the power industry according to claim 1, characterized in that, In step 2), the identification of position-change atomic components is used to determine the spatiotemporal relationship of a target at different times. The specific steps are as follows: Step A): First, set a target's position change distance threshold. and position change angle threshold ; Step B): In Time's up At any given moment, locate the position of a target within the field of view in space, where the positions within the field of view are respectively... Coordinates of the predicted bounding box of a target at time 1 ,as well as Coordinates of the predicted bounding box of a target at time 1 ; Step C): Calculate the Euclidean distance between the two coordinate points. Specifically, it is as follows: (5) (5); Step D): Select The axis is horizontal; calculate the angle between two coordinate points. Specifically, it is the following formula (6): (6); Step E): Calculate the recognition result of the spatiotemporal relationship of a target at different times. The formula is as follows (7): (7); in, and ,but A value of 1 indicates successful recognition; otherwise, If the value is 0, recognition failed.

6. The multimodal AI method for job assessment in the power industry according to claim 1, characterized in that, In step 3), the score of the voice component is recorded in the exam questions, wherein the score of the voice component in the exam questions is recorded as follows: .

7. The multimodal AI method for job assessment in the power industry according to claim 1, characterized in that, In step 4), the score for the test questions is calculated using the following steps: Step I): The score of the visual component is calculated as the confidence level of the image or video segment recognized by the SAM model. Multiply by the allocation of visual components ; Step II): The score of the speech component is calculated as the sum of the weights of the keywords in the language text recognized by the ASR model. Multiply by the allocation of voice-related components ;in, For the identified first One keyword, The total number of keywords identified; Step III): Final score of the exam questions The calculation method is as follows, as shown in formula (8): (8)； in, For the first Visual-related test points To identify the first Confidence level of images or video clips for each visual test point , For the first The allocation of visual components for each visual test point , For a test question there are Visual-related test points for The score of visual components for each visual test point; For the first One key point related to pronunciation. For the first The allocation of points for speech-related components in each speech-related test point , For a test question there are One key point related to pronunciation. for The score of the voice component for each voice-related test point.

8. A recognition system for a multimodal AI method applied to job assessment in the power industry, used to execute the multimodal AI method for job assessment in the power industry according to any one of claims 1-7, characterized in that, include: SAM model, used for image or video recognition; The ASR model is used to recognize and convert the speech expressed by test takers into written language. The performance evaluation module is used for job performance evaluation. The assessment and evaluation module is connected to the SAM model and the ASR model. The image or video recognition data in the SAM model and the language and text recognition data in the ASR model are sent to the assessment and evaluation module, which then assesses and evaluates the job.

9. The recognition system of the multimodal AI method for job assessment in the power industry according to claim 8, characterized in that, The SAM model includes: a target recognition atomic component module for identifying the spatial relationship between two targets in an image or video; a recognition atomic component module for identifying two label states of a target at different time points; and a position change atomic component module for identifying and judging the spatiotemporal relationship of a target at different times. The ASR model includes: a speech-related component module for converting speech into written text for recognition; The assessment and evaluation module is connected to the target recognition atomic component module, the atomic component module, and the position change atomic component module, and the assessment and evaluation module is connected to the voice component module; The assessment and evaluation module combines visual components (a combination of target recognition atomic components, atomic components, and position change atomic components) with voice components from the voice component module for job assessment and evaluation.

Images