Small Kernel Research Articles

The Effect of Kernel Size on Bioactive Compounds in Kalınkara Hazelnut Cultivar

Objective: This study was carried out to determine the effect of kernel size on the change of bioactive compounds in Kalınkara hazelnut cultivar.
 Materials and Methods: This study was conducted on the Kalınkara hazelnut cultivar grown in a producer's orchard in Ulubey (Ordu) district. The study was designed according to the randomized plot design with three replications and three ocak in each replication. Harvested nuts were classified according to kernel size as “small”, “medium” and “large”. Then, total phenolics, total flavonoids, and antioxidant activity were determined. In addition, correlation coefficients were calculated, and principal component analyses were performed to determine the relationships between the examined features.
 Results: In the study, it was observed that kernel size was effective on bioactive compounds in Kalınkara hazelnut cultivar. According to the research findings, the total phenolics was determined between 759 mg 100 g-1 (medium) and 819 mg 100 g-1 (large), the total flavonoids was determined between 8.2 mg 100 g-1 (medium) and 8.7 mg 100 g-1 (small), and total antioxidant activity was determined between 0.46 mmol 100 g-1 (medium) and 0.60 mmol 100 g-1 (large) according to the FRAP assay, while it was determined between 1.83 mmol 100 g-1 (medium) and 1.92 mmol 100 g-1 (small) according to the DPPH assay. In addition, a positive correlation was determined between the total phenolics and antioxidant activity. Accordingly, while the correlation coefficient between total phenolics and DPPH assay was r=0.921***, it was r=0.982*** with FRAP assay.
 Conclusion: As a result of the research, it was determined that the total flavonoids, and antioxidant activity according to the DPPH test were higher in small kernels. In comparison, total phenolics, and antioxidant activity according to the FRAP assay were higher in large kernels.

Face expression recognition based on residual network and support vector machine

In recent years, Facial expression recognition (FER) has become an outstanding research field, which content academia and industry alike. In order to alleviate the gradient disappearance in facial expression recognition caused by increasing network depth, this paper proposes joint residual network and support vector machine (SVM) to build a model. The algorithm uses a small convolutional kernel and a deep network structure. In addition, a multi-task cascaded convolutional network (MTCNN) is constructed for data pre-processing. Meanwhile, migration learning is introduced to overcome the shortcomings of insufficient data and prevent overfitting. Extensive experiments are conducted on the FER-2013 dataset. The model achieves a recognition rate of 98.6% in the test, meanwhile the accuracy of SVMs category is about 1% better than that of SoftMax. The experimental results proof that the model hereby propose outperforms different networks and previous methods. It further perceives that the method hereby propose is effective in recognizing facial expressions.

Hybrid CNN-transformer network for interactive learning of challenging musculoskeletal images

Background and objectivesSegmentation of regions of interest (ROIs) such as tumors and bones plays an essential role in the analysis of musculoskeletal (MSK) images. Segmentation results can help with orthopaedic surgeons in surgical outcomes assessment and patient's gait cycle simulation. Deep learning-based automatic segmentation methods, particularly those using fully convolutional networks (FCNs), are considered as the state-of-the-art. However, in scenarios where the training data is insufficient to account for all the variations in ROIs, these methods struggle to segment the challenging ROIs that with less common image characteristics. Such characteristics might include low contrast to the background, inhomogeneous textures, and fuzzy boundaries. Methodswe propose a hybrid convolutional neural network – transformer network (HCTN) for semi-automatic segmentation to overcome the limitations of segmenting challenging MSK images. Specifically, we propose to fuse user-inputs (manual, e.g., mouse clicks) with high-level semantic image features derived from the neural network (automatic) where the user-inputs are used in an interactive training for uncommon image characteristics. In addition, we propose to leverage the transformer network (TN) – a deep learning model designed for handling sequence data, in together with features derived from FCNs for segmentation; this addresses the limitation of FCNs that can only operate on small kernels, which tends to dismiss global context and only focus on local patterns. ResultsWe purposely selected three MSK imaging datasets covering a variety of structures to evaluate the generalizability of the proposed method. Our semi-automatic HCTN method achieved a dice coefficient score (DSC) of 88.46 ± 9.41 for segmenting the soft-tissue sarcoma tumors from magnetic resonance (MR) images, 73.32 ± 11.97 for segmenting the osteosarcoma tumors from MR images and 93.93 ± 1.84 for segmenting the clavicle bones from chest radiographs. When compared to the current state-of-the-art automatic segmentation method, our HCTN method is 11.7%, 19.11% and 7.36% higher in DSC on the three datasets, respectively. ConclusionOur experimental results demonstrate that HCTN achieved more generalizable results than the current methods, especially with challenging MSK studies.

A Novel Deep Convolutional Neural Network Combining Global Feature Extraction and Detailed Feature Extraction for Bearing Compound Fault Diagnosis.

This study researched the application of a convolutional neural network (CNN) to a bearing compound fault diagnosis. The proposed idea lies in the ability of CNN to automatically extract fault features from complex raw signals. In our approach, to extract more effective features from a raw signal, a novel deep convolutional neural network combining global feature extraction with detailed feature extraction (GDDCNN) is proposed. First, wide and small kernel sizes are separately adopted in shallow and deep convolutional layers to extract global and detailed features. Then, the modified activation layer with a concatenated rectified linear unit (CReLU) is added following the shallow convolution layer to improve the utilization of shallow global features of the network. Finally, to acquire more robust features, another strategy involving the GMP layer is utilized, which replaces the traditional fully connected layer. The performance of the obtained diagnosis was validated on two bearing datasets. The results show that the accuracy of the compound fault diagnosis is over 98%. Compared with three other CNN-based methods, the proposed model demonstrates better stability.

Respiratory Exertion Using Virtual Reality

Magnetic Resonance Imaging (MRI) is a widely used imaging technique to assess these tumors. The large amount of data produced by MRI prevents manual segmentation in a reasonable time. Automatic and reliable segmentation methods are required. Independent projection-based classification (LIPC) is used to segment the tumor region. Here, also investigated the use of intensity normalization as a pre-processing step, which though not common in CNN-based segmentation methods, proved together with data augmentation to be very effective for brain tumor segmentation in MRI images. Here, proposed an automatic segmentation method based on Convolutional Neural Networks (CNN), exploring small 3x3 kernels. The use of small kernels allows designing a deeper architecture, besides having a positive effect against over fitting, given the fewer number of weights in the network. Results reported on public benchmarks reveal that our architecture achieves competitive accuracy compared to the state- of-the-art brain tumor segmentation methods while being computationally efficient.. Index Terms—Digital Image Processing,Positron Emission Tomography,PhotomultiplierTube,ComputedTomography,Magne tic Resonance Image,Fluid-Attenuated Inversion Recovery ,Single Photon Emission Computed Tomography TR Repetition Time, echo time ,Region of Interest.

Single-Slope ADC With Embedded Convolution Filter for Global-Shutter CMOS Image Sensors

This paper presents an analog-to-digital converter (ADC) suitable for acquisition and processing of images in the global-shutter mode at the pixel level. The ADC consists of an analog comparator, a multi-directional shift register for the comparator states, and a 16-bit reversible binary counter with programmable step size. It works in the traditional single-slope mode. The novelty is that during each step of the reference ramp, neighboring pixels can exchange status information. During the conversion, the direction and step size of the counter are set globally to realize the corresponding coefficient of a convolution kernel. This technique does not slow down the conversion when used for small kernels (3W3) and does not significantly increase sensor noise. Convolution windows of arbitrary size can be implemented. The concept was verified in an experimental 64W64 imaging array implemented in 180 nm CMOS technology.

Entangled kernel of coded information using quantum string

The possibility of encoding classical string of information into a quantum string via CNOT operation and Hadamard gate is discussed. It has been shown that maximizing/minimizing the quantum correlation of the generated quantum string depends on the values of the initial classical strings. The kernel of each string, as well as, the total string had been examined, where by applying the Hadamard gate on the classical string, one can decrease its similarity as well as the similarity of the total kernel. However, in the absence of the Hadamard gate, the large similarity of the final quantum string has been observed. The kernel of the final quantum string may be used as an indictor of the security degree of the encoded information, where the large degree of security has depicted at small kernel. The high degree of security can be implemented by applying the Hadamard gate on only one classical string.

A Small Object Detection Method for Oil Leakage Defects in Substations Based on Improved Faster-RCNN.

Since substations are key parts of power transmission, ensuring the safety of substations involves monitoring whether the substation equipment is in a normal state. Oil leakage detection is one of the necessary daily tasks of substation inspection robots, which can immediately find out whether there is oil leakage in the equipment in operation so as to ensure the service life of the equipment and maintain the safe and stable operation of the system. At present, there are still some challenges in oil leakage detection in substation equipment: there is a lack of a more accurate method of detecting oil leakage in small objects, and there is no combination of intelligent inspection robots to assist substation inspection workers in judging oil leakage accidents. To address these issues, this paper proposes a small object detection method for oil leakage defects in substations. This paper proposes a small object detection method for oil leakage defects in substations, which is based on the feature extraction network Resnet-101 of the Faster-RCNN model for improvement. In order to decrease the loss of information in the original image, especially for small objects, this method is developed by canceling the downsampling operation and replacing the large convolutional kernel with a small convolutional kernel. In addition, the method proposed in this paper is combined with an intelligent inspection robot, and an oil leakage decision-making scheme is designed, which can provide substation equipment oil leakage maintenance recommendations for substation workers to deal with oil leakage accidents. Finally, the experimental validation of real substation oil leakage image collection is carried out by the intelligent inspection robot equipped with a camera. The experimental results show that the proposed FRRNet101-c model in this paper has the best performance for oil leakage detection in substation equipment compared with several baseline models, improving the Mean Average Precision (mAP) by 6.3%, especially in detecting small objects, which has improved by 12%.

HyFormer: Hybrid Grouping-Aggregation Transformer and Wide-Spanning CNN for Hyperspectral Image Super-Resolution

Hyperspectral image (HSI) super-resolution is a practical and challenging task as it requires the reconstruction of a large number of spectral bands. Achieving excellent reconstruction results can greatly benefit subsequent downstream tasks. The current mainstream hyperspectral super-resolution methods mainly utilize 3D convolutional neural networks (3D CNN) for design. However, the commonly used small kernel size in 3D CNN limits the model’s receptive field, preventing it from considering a wider range of contextual information. Though the receptive field could be expanded by enlarging the kernel size, it results in a dramatic increase in model parameters. Furthermore, the popular vision transformers designed for natural images are not suitable for processing HSI. This is because HSI exhibits sparsity in the spatial domain, which can lead to significant computational resource waste when using self-attention. In this paper, we design a hybrid architecture called HyFormer, which combines the strengths of CNN and transformer for hyperspectral super-resolution. The transformer branch enables intra-spectra interaction to capture fine-grained contextual details at each specific wavelength. Meanwhile, the CNN branch facilitates efficient inter-spectra feature extraction among different wavelengths while maintaining a large receptive field. Specifically, in the transformer branch, we propose a novel Grouping-Aggregation transformer (GAT), comprising grouping self-attention (GSA) and aggregation self-attention (ASA). The GSA is employed to extract diverse fine-grained features of targets, while the ASA facilitates interaction among heterogeneous textures allocated to different channels. In the CNN branch, we propose a Wide-Spanning Separable 3D Attention (WSSA) to enlarge the receptive field while keeping a low parameter number. Building upon WSSA, we construct a wide-spanning CNN module to efficiently extract inter-spectra features. Extensive experiments demonstrate the superior performance of our HyFormer.

ZmBELL10 interacts with other ZmBELLs and recognizes specific motifs for transcriptional activation to modulate internode patterning in maize.

Plant height is an important agronomic trait that affects crop yield. Elucidating the molecular mechanism underlying plant height regulation is also an important question in developmental biology. Here, we report that a BELL transcription factor, ZmBELL10, positively regulates plant height in maize (Zea mays). Loss of ZmBELL10 function resulted in shorter internodes, fewer nodes, and smaller kernels, while ZmBELL10 overexpression increased plant height and hundred-kernel weight. Transcriptome analysis and chromatin immunoprecipitation followed by sequencing showed that ZmBELL10 recognizes specific sequences in the promoter of its target genes and activates cell division- and cell elongation-related gene expression, thereby influencing node number and internode length in maize. ZmBELL10 interacted with several other ZmBELL proteins via a spatial structure in its POX domain to form protein complexes involving ZmBELL10. All interacting proteins recognized the same DNA sequences, and their interaction with ZmBELL10 increased target gene expression. We identified the key residues in the POX domain of ZmBELL10 responsible for its protein-protein interactions, but these residues did not affect its transactivation activity. Collectively, our findings shed light on the functions of ZmBELL10 protein complexes and provide potential targets for improving plant architecture and yield in maize.

Seismic velocity inversion transformer

Velocity model inversion is one of the most challenging tasks in seismic exploration, and an accurate velocity model is essential for high-resolution seismic imaging. Recently, velocity inversion methods based on deep learning (DL), particularly convolutional neural networks (CNNs), have attracted considerable attention from the seismic exploration community. These researchers aim to directly estimate the velocity model from raw seismograms using a well-trained model. Although CNN-based velocity inversion methods have demonstrated remarkable performance in terms of intelligence and automation, their inversion performance is often constrained by a limited long-range dependence. Specifically, when conducting a convolutional operation on raw seismic data using small kernels (i.e., 1 × 1, 3 × 3, 5 × 5, and 7 × 7), CNN-based methods extract only the local features and neglect the weak spatial correlation between different local features that reflect the information of the same interface. This correlation could assist CNN in providing an overview of the seismic data and promote inversion performance when using DL. Furthermore, the time-varying properties of seismic data pose a challenge to the weight sharing of CNNs. Here, we have developed a new DL framework based on a transformer, called the seismic velocity inversion transformer (SVIT), to address the problem of velocity inversion. SVIT uses a self-attention mechanism to capture the long-range dependence of seismic data, rather than stacking multiple convolutional layers as in CNNs. Thus, SVIT can provide more informative remote features for building velocity models. The validity and reliability of the proposed method are demonstrated through numerical experiments using synthetic models. Compared with the conventional full-waveform inversion method and an existing CNN-based velocity inversion method, our SVIT indicates greater consistency with the target in terms of the velocity value, subsurface structures, and geologic interfaces and is expected to provide a new DL-based solution to resolve inversion problems.

LKC-Net: large kernel convolution object detection network

Deep learning-based object detection methods have achieved great performance improvement. However, since small kernel convolution has been widely used, the semantic feature is difficult to obtain due to the small receptive fields, and the key information cannot be highlighted, resulting in a series of problems such as wrong detection, missing detection, and repeated detection. To overcome these problems, we propose a large kernel convolution object detection network based on feature capture enhancement and vast receptive field attention, called LKC-Net. Firstly, a feature capture enhancement block based on large kernel convolution is proposed to improve the semantic feature capturing ability, and depth convolution is used to reduce the number of parameters. Then, the vast receptive filed attention mechanism is constructed to enhance channel direction information extraction ability, and it is more compatible with the proposed backbone than other existing attention mechanisms. Finally, the loss function is improved by introducing the SIoU, which can overcome the angle mismatch problem between the ground truth and prediction box. Experiments are conducted on Pascal VOC and MS COCO datasets for demonstrating the performance of LKC-Net.

ZmDRR206 Regulates Nutrient Accumulation in Endosperm through Its Role in Cell Wall Biogenesis during Maize Kernel Development.

Dirigent proteins (DIRs) contribute to plant fitness by dynamically reorganizing the cell wall and/or by generating defense compounds during plant growth, development, and interactions with environmental stresses. ZmDRR206 is a maize DIR, it plays a role in maintaining cell wall integrity during seedling growth and defense response in maize, but its role in regulating maize kernel development is unclear. Association analysis of candidate genes indicated that the natural variations of ZmDRR206 were significantly associated with maize hundred-kernel weight (HKW). ZmDRR206 plays a dominant role in storage nutrient accumulation in endosperm during maize kernel development, ZmDRR206 overexpression resulted in small and shrunken maize kernel with significantly reduced starch content and significantly decreased HKW. Cytological characterization of the developing maize kernels revealed that ZmDRR206 overexpression induced dysfunctional basal endosperm transfer layer (BETL) cells, which were shorter with less wall ingrowth, and defense response was constitutively activated in developing maize kernel at 15 and 18 DAP by ZmDRR206 overexpression. The BETL-development-related genes and auxin signal-related genes were down-regulated, while cell wall biogenesis-related genes were up-regulated in developing BETL of the ZmDRR206-overexpressing kernel. Moreover, the developing ZmDRR206-overexpressing kernel had significantly reduced contents of the cell wall components such as cellulose and acid soluble lignin. These results suggest that ZmDRR206 may play a regulatory role in coordinating cell development, storage nutrient metabolism, and stress responses during maize kernel development through its role in cell wall biogenesis and defense response, and provides new insights into understanding the mechanisms of kernel development in maize.

Human Activity Recognition in a Realistic and Multiview Environment Based on Two-Dimensional Convolutional Neural Network

Recognition of human activity based on convolutional neural network has received the interest of researchers in recent years due to its significant improvement in accuracy. A large number of algorithms based on the deep learning approach have been proposed for activity recognition purpose. However, with the increasing advancements in technologies having limited computational resources, it needs to design an efficient deep learning-based approaches with improved utilization of computational resources. This paper presents a simple and efficient 2-Dimensional convolutional neural network (2-D CNN) architecture with very small size convolutional kernel for human activity recognition. The merit of the proposed CNN architecture over standard deep learning architectures is fewer trainable parameters and lesser memory requirement which enables it to train the proposed CNN architecture on low GPU memory-based devices and also works well with smaller as well as larger size datasets. The proposed approach consists of mainly four stages: namely (1) creation of dataset and data augmentation, (2) designing 2-D convolutional neural network (CNN) architecture, (3) the proposed 2-D CNN architecture trained from scratch up to optimum stage, and (4) evaluation of the trained 2-D CNN architecture. To illustrate the effectiveness of the proposed architecture several extensive experiments are conducted on three publicly available datasets, namely IXMAS, YouTube, and UCF101 dataset. The results of the proposed method and its comparison with other state-of-the-art methods [8-12,14,18-26,29-33] demonstrate the usefulness of the proposed method.

Segmentation of prostate region in magnetic resonance images based on improved V-Net

Magnetic resonance (MR) imaging is an important tool for prostate cancer diagnosis, and accurate segmentation of MR prostate regions by computer-aided diagnostic techniques is important for the diagnosis of prostate cancer. In this paper, we propose an improved end-to-end three-dimensional image segmentation network using a deep learning approach to the traditional V-Net network (V-Net) network in order to provide more accurate image segmentation results. Firstly, we fused the soft attention mechanism into the traditional V-Net's jump connection, and combined short jump connection and small convolutional kernel to further improve the network segmentation accuracy. Then the prostate region was segmented using the Prostate MR Image Segmentation 2012 (PROMISE 12) challenge dataset, and the model was evaluated using the dice similarity coefficient (DSC) and Hausdorff distance (HD). The DSC and HD values of the segmented model could reach 0.903 and 3.912 mm, respectively. The experimental results show that the algorithm in this paper can provide more accurate three-dimensional segmentation results, which can accurately and efficiently segment prostate MR images and provide a reliable basis for clinical diagnosis and treatment.

A new deep convolutional neural network incorporating attentional mechanisms for ECG emotion recognition

Using ECG signals captured by wearable devices for emotion recognition is a feasible solution. We propose a deep convolutional neural network incorporating attentional mechanisms for ECG emotion recognition. In order to address the problem of individuality differences in emotion recognition tasks, we incorporate an improved Convolutional Block Attention Module (CBAM) into the proposed deep convolutional neural network. The deep convolutional neural network is responsible for capturing ECG features. Channel attention in CBAM is responsible for adding weight information to ECG features of different channels and spatial attention is responsible for the weighted representation of ECG features of different regions inside the channel. We used three publicly available datasets, WESAD, DREAMER, and ASCERTAIN, for the ECG emotion recognition task. The new state-of-the-art results are set in three datasets for multi-class classification results, WESAD for tri-class results, and ASCERTAIN for two-category results, respectively. A large number of experiments are performed, providing an interesting analysis of the design of the convolutional structure parameters and the role of the attention mechanism used. We propose to use large convolutional kernels to improve the effective perceptual field of the model and thus fully capture the ECG signal features, which achieves better performance compared to the commonly used small kernels. In addition, channel attention and spatial attention were added to the deep convolutional model separately to explore their contribution levels. We found that in most cases, channel attention contributed to the model at a higher level than spatial attention.

Microstructural evolution, mechanical properties and surface quality of TC11 titanium alloy subjected to waterjet-assisted laser direct inscription

Waterjet-assisted laser direct inscription (WJALDI) technology has presented a positive effect to alleviate the thermal damage in workpieces caused by laser ablation. This research pays attention to the effect of aqueous media on the microstructure evolution and phase composition of materials during laser ablation, and few relevant researches have been reported. In this work, titanium alloy (TC11) was applied to fabricate microgrooves utilizing WJALDI and laser direct inscription in the air (LDI). The interaction effects between the laser, water, and the material within a complete ablation cycle were analyzed in depth. The geometry and thermal damage region were observed and discussed to prove the machining performance. The microstructure evolution that happened on and underneath the microgroove surface was also studied. The results revealed that the effect of cooling and impacting significantly reduced the accumulation of debris and molten metal, and better processing quality was obtained in WJALDI. According to the analysis of electron backscattered diffraction (EBSD), the grain refinement of the inner material was alleviated, low-angle grain boundaries with the features of uniform distribution were generated, and the adjacent grains of α phase have small kernel average misorientation. In addition, compared with LDI, WJALDI offered a wide microgroove with a small size of heat affected zone (HAZ) and thin thickness of the recast layer, while less depth. Therefore, the developed technology in this work could be a promising approach for the micromachining of titanium alloy or even other difficult-to-cut materials with high quality.

Motif-GCNs With Local and Non-Local Temporal Blocks for Skeleton-Based Action Recognition.

Recent works have achieved remarkable performance for action recognition with human skeletal data by utilizing graph convolutional models. Existing models mainly focus on developing graph convolutional operations to encode structural properties of a skeletal graph, whose topology is manually predefined and fixed over all action samples. Some recent works further take sample-dependent relationships among joints into consideration. However, the complex relationships between arbitrary pairwise joints are difficult to learn and the temporal features between frames are not fully exploited by simply using traditional convolutions with small local kernels. In this paper, we propose a motif-based graph convolution method, which makes use of sample-dependent latent relations among non-physically connected joints to impose a high-order locality and assigns different semantic roles to physical neighbors of a joint to encode hierarchical structures. Furthermore, we propose a sparsity-promoting loss function to learn a sparse motif adjacency matrix for latent dependencies in non-physical connections. For extracting effective temporal information, we propose an efficient local temporal block. It adopts partial dense connections to reuse temporal features in local time windows, and enrich a variety of information flow by gradient combination. In addition, we introduce a non-local temporal block to capture global dependencies among frames. Our model can capture local and non-local relationships both spatially and temporally, by integrating the local and non-local temporal blocks into the sparse motif-based graph convolutional networks (SMotif-GCNs). Comprehensive experiments on four large-scale datasets show that our model outperforms the state-of-the-art methods. Our code is publicly available at https://github.com/wenyh1616/SAMotif-GCN.

Monocular Depth Estimation Network Based on Swin Transformer

Estimating depth from a single image is challenging because a single 2D image may correspond to many different 3D scenes with the same depth. While most deep learning based depth prediction methods extract depth features using small convolutional kernels with small receptive fields, which results in deformed depth edges and inaccurate depth values of distant objects in the depth estimation results. Aiming at this problem, we propose a depth estimation network based on Swin Transformer and the encoder-decoder structure. We construct the encoder using the Swin Transformer network, which can encode long-range spatial dependency and extract features on various scales and across different channels. The decoder of the proposed network is in charge of fusing the features from the encoder by the operations of interpolation, concatenation, and convolution. Experiments on KITTI and NYUv2 datasets show that our proposed network can get more accurate depth edges and depth values than the state-of-the-art methods.

Tetratricopeptide-containing SMALL KERNEL 11 is essential for the assembly of cytochrome c oxidase in maize mitochondria.

Assembly of the functional complexes of the mitochondrial respiratory chain requires sophisticated and efficient regulatory mechanisms. In plants, the subunit composition and assembly factors involved in the biogenesis of cytochrome c oxidase (complex IV) are substantially less defined than in mammals and yeast. In this study, we cloned maize (Zea mays) Small kernel 11 (Smk11) via map-based cloning. Smk11 encodes a mitochondria-localized tetratricopeptide repeat protein. Disruption of Smk11 severely affected the assembly and activity of mitochondrial complex IV, leading to delayed plant growth and seed development. Protein interactions studies revealed that SMK11 might interact with four putative complex IV assembly factors, Inner membrane peptidase 1A (ZmIMP1A), MYB domain protein 3R3 (ZmMYB3R-3), cytochrome c oxidase 23 (ZmCOX23), and mitochondrial ferredoxin 1 (ZmMFDX1), among which ZmMFDX1 might interact with subunits ZmCOX6a and ZmCOX-X1; ZmMYB3R-3 might also interact with ZmCOX6a. The mutation of SMK11 perturbed the normal assembly of these subunits, leading to the inactivation of complex IV. The results of this study revealed that SMK11 serves as an accessory assembly factor required for the normal assembly of subunits into complex IV, which will accelerate the elucidation of the assembly of complex IV in plant mitochondria.